# AI Policy Observatory - all country summaries > See https://aipolicy.ius.uzh.ch/llms.txt for the per-country index. # AI Policy: Australia (AU) Canonical page: https://aipolicy.ius.uzh.ch/compare/australia Last updated: 2026-03-02T15:20:12.473Z ## General Description Australia has a "[National AI Plan"](https://www.industry.gov.au/publications/national-ai-plan) with the goals of (1) capturing the opportunities presented by AI and building domestic capacity; (2) spreading the benefits through widespread adoption, support and training for workers and improved public services; and (3) keeping Australians safe through legislative and regulatory frameworks that mitigate AI harms while promoting responsible practices and international engagement that upholds Australia’s values. The third element of this strategy involves: - Establishing an AI Safety Institute - Clarifying consumer protection laws in consultation with states and territories - Enforceable industry codes under the Online Safety Act 2021 (Cth), supplemented with further reforms to tackle ‘nudify’ apps and algorithmic bias - Reforms to copyright law - Reviewing AI regulation in medical device software - Various measures in the context of AI security - Updating Australia’s privacy laws, particularly the Privacy Act 1988 (Cth) - Providing guidance on responsible AI adoption, including recommended transparency measures and specific guidance in regulated sectors - Governance of government AI use - Guidance for AI in schools - Participating in AI standards development internationally Australia has a federal system and some states have also looked at the sufficiency of legal and regulatory frameworks in their own jurisdictions. There has been a [parliamentary inquiry](https://www.parliament.nsw.gov.au/committees/inquiries/Pages/inquiry-details.aspx?pk=2968) on the issue in New South Wales and a [Select Committee on Artificial Intelligence in South Australia](https://www.parliament.sa.gov.au/en/News/2023/07/11/03/37/Select-Committee-on-Artificial-Intelligence). Some states have also issued directives or consulted on issues surrounding the use of artificial intelligence in courts and tribunals. ## Steps Taken The question of how AI would be regulated in Australia was the subject of significant analysis. In September 2024, the Australian government articulated proposals for introducing mandatory guardrails for AI in high-risk settings as part of its “safe and responsible AI” project. There were ten proposed mandatory guardrails in relation to which the government is considering three regulatory options, namely: Option 1: Adapt existing regulatory frameworks to include the proposed mandatory guardrails Option 2: Introduce framework legislation, with associated amendments to existing legislation Option 3: Introduce a new cross-economy AI Act. Following its consultations on that proposal, the government instead opted for the National AI Plan, as described above, on 2 December 2025. Most of the tasks are still in development. What exists currently are: 1. a Voluntary AI Safety Standard in collaboration with the National Artificial Intelligence Centre, setting out how to safely and responsible use and innovate with artificial intelligence; 2. an Australian Framework for Generative Artificial Intelligence (AI) in Schools; and 3. a national framework for the assurance of AI in government (in partnership with state and territory governments). The Australian Government is also working on a number of related initiatives including: 1. [privacy law reform](https://www.ag.gov.au/integrity/consultations/review-privacy-act-1988); 2. [reforms](https://www.pmc.gov.au/resources/government-response-royal-commission-robodebt-scheme) associated with the Royal Commission into the Robodebt Scheme; 3. [statutory review](https://www.infrastructure.gov.au/department/media/publications/report-statutory-review-online-safety-act-2021) of the Online Safety Act 2021; 4. a [regulatory framework](https://www.ntc.gov.au/transport-reform/automated-vehicle-program) for automated vehicles; 5. establishing a [Copyright and Artificial Intelligence Reference Group](https://www.ag.gov.au/rights-and-protections/copyright/copyright-and-artificial-intelligence-reference-group-cairg). Internationally, Australia has endorsed the Bletchley Declaration, the Seoul Declaration and the Paris Statement. It participates in the UN Global Digital Compact, the Hiroshima AI Process and the Global Partnership on AI. [](https://www.industry.gov.au/publications/bletchley-declaration-countries-attending-ai-safety-summit-1-2-november-2023) ## Approaches Australia has considered and rejected a proposal for technology-specific laws, instead focusing on law reform, updating existing laws and guidelines. As a country that imports AI systems and components, the Australian government is also focussed on contributing to global governance and standards for AI as well as interoperability with other countries’ laws. ## Content Following a decision not to enact legislation focussing specifically on AI, there are instead a variety of relevant laws some of which may be updated to manage new risks associated with AI. There are various forms of soft law that focus specifically on AI, including the Voluntary AI Safety Standard, the Australian Framework for Generative Artificial Intelligence (AI) in Schools, and the National framework for the assurance of artificial intelligence in government. ## Sources - Report of the Statutory Review of the Online Safety Act 2021, 4 February 2025: https://www.infrastructure.gov.au/department/media/publications/report-statutory-review-online-safety-act-2021 - Proposals paper: Introducing mandatory guardrails for AI in high-risk settings, September 2024: https://consult.industry.gov.au/ai-mandatory-guardrails - National framework for the assurance of artificial intelligence in government, 21 June 2024: https://www.finance.gov.au/government/public-data/data-and-digital-ministers-meeting/national-framework-assurance-artificial-intelligence-government - Regulatory Framework for automated vehicles, April 2024: https://www.ntc.gov.au/transport-reform/automated-vehicle-program - Consultation and interim response on Safe and Responsible AI, 17 January 2024: https://consult.industry.gov.au/supporting-responsible-ai - CAIRG – Copyright and Artificial Intelligence Reference Group, established 5 December 2023: https://www.ag.gov.au/rights-and-protections/copyright/copyright-and-artificial-intelligence-reference-group-cairg - Australian Framework for Generative Artificial Intelligence (AI) in Schools, 17 November 2023: https://www.education.gov.au/schooling/resources/australian-framework-generative-artificial-intelligence-ai-schools - Bletchley Declaration, 2 November 2023: https://www.industry.gov.au/publications/bletchley-declaration-countries-attending-ai-safety-summit-1-2-november-2023 - Government response to the Royal Commission into the Robodebt Scheme, November 2023: https://www.pmc.gov.au/resources/government-response-royal-commission-robodebt-scheme - Voluntary AI Safety Standard, 5 September 2024: https://www.industry.gov.au/publications/voluntary-ai-safety-standard - Inquiry into artificial intelligence (AI) in New South Wales, 27 June 2023: https://www.parliament.nsw.gov.au/committees/inquiries/Pages/inquiry-details.aspx?pk=2968 - Review of the Privacy Act 1988, 16 February 2023: https://www.ag.gov.au/integrity/consultations/review-privacy-act-1988 --- # AI Policy: Brazil (BR) Canonical page: https://aipolicy.ius.uzh.ch/compare/brazil Last updated: 2025-06-30T07:59:18.290Z ## General Description The Brazilian Congress (*Congresso Nacional*) is discussing the adoption of a comprehensive AI regulation ([Bill nº 2.338/2023](https://www25.senado.leg.br/web/atividade/materias/-/materia/157233)), which was presented by the Senate's President as an alternative for other bills on the subject. Its purpose is to establish general rules for the development, implementation, and use of AI systems in Brazil. The Bill was heavily influenced by the European Union AI Act and is centred around risk assessment and mitigation. After it was approved by the Senate in late 2024, the Bill is now moving on to the House of Representatives for a second vote. Aside from defining how AI systems should be classified according to the risks they pose, the Bill, unlike the AI Act, grants a number of individual rights to the people "affected" by AI applications and defines how AI governance should look like in the country, including the designation of a "competent authority" by the Executive to oversee AI regulation in Brazil. ## Steps Taken - In December of 2024, the Brazilian Senate approved [Bill nº 2.338/2023](https://www25.senado.leg.br/web/atividade/materias/-/materia/157233), which will now move on to the House of Representatives for a second vote. - It is important to note that it is just a Bill that still needs to be voted by both houses of Congress and sanctioned by the President before going into force. Given that the Bill is still pending a final vote, it is reasonable to expect significant changes to the text before a final version can be agreed upon, specially in the House of Representatives. The Bill also has a vacatio legis provision of one year, meaning it won't go into effect until one year after it is sanctioned by the President. - In July of 2024, the Ministry of Science, Technology and Innovation published the new Brazilian Artificial Intelligence Plan ([*Plano Brasileiro de Inteligência Artificial - PBIA*](https://www.gov.br/lncc/pt-br/assuntos/noticias/ultimas-noticias-1/plano-brasileiro-de-inteligencia-artificial-pbia-2024-2028)). With a planned investment of R$23 billion over four years, the ambitious plan aims to transform the country into a world reference in the use of AI, especially in the public sector. - In February of 2024, the President of the Senate publicly stated that AI regulation is a priority and that [Bill nº 2.338/2023](https://www25.senado.leg.br/web/atividade/materias/-/materia/157233) should be voted as soon as possible. - In May of 2023, the President of the Senate officially presented Bill nº 2.338/2023 to the floor, which is based on the commission's final report. - In December of 2022, the commission published its [final report.](https://www.stj.jus.br/sites/portalp/SiteAssets/documentos/noticias/Relato%CC%81rio%20final%20CJSUBIA.pdf) - In 2022, the President of the Senate, Senator Rodrigo Pacheco, mandated the creation of a "special commission of jurists" in charge of drafting a single alternative to the other bills. - In July of 2021, the Ministry of Science, Technology and Innovation published the "Brazilian Artificial Intelligence Strategy" ([*Estratégia Brasileira de Inteligência Artificial - EBIA*](https://www.gov.br/mcti/pt-br/acompanhe-o-mcti/transformacaodigital/arquivosinteligenciaartificial/ebia-documento_referencia_4-979_2021.pdf)). - Between 2019 and 2021, the first legislative bills on AI were proposed by representatives in the National Congress, including [Bill nº 21/2020](https://www.congressonacional.leg.br/materias/materias-bicamerais/-/ver/pl-21-2020#:~:text=Ementa%3A,Brasil%3B%20e%20d%C3%A1%20outras%20provid%C3%AAncias.) on the use of AI in Brazil. ## Approaches Brazil's approach to AI regulation is centred around risk assessment and mitigation. Every AI system should undergo a preliminary assessment to determine if it falls within the scope of what the Bill considers to be a high-risk application. If that is the case, the competent authority (which is still to be defined) will be notified and an "algorithmic impact assessment" will be conducted under its guidance. Unlike the AI Act in the EU – which mentions four levels of risk for AI systems –, the Brazilian Bill only mentions applications that present an (a) excessive or (b) high risk – not to mention a third, presumed level of risk for those systems that do not fit any of the two previous categories. According to the Bill, AI systems that pose excessive risks should be prohibited and, therefore, cannot be implemented in the country. This includes the use of AI for social credit scoring and AI applications that may explore any social, economic or personal vulnerabilities to its advantage. The other cases listed by the Bill are considered to be high-risk, including AI for criminal investigation, biometric identification, medical diagnosis, (financial) credit scoring, autonomous vehicles, critical infrastructure, job recruiting, education, etc. The competent authority will have the ability to revise the list of AI systems that present an excessive or high risk in the future. ## Content One of the main criticisms that the Bill is facing is its overreliance on the same approach as the AI Act. One cannot lose sight of the fact that the AI Act is part of a broader European regulatory framework that includes initiatives as diverse (and complementary) as the Digital Services Act, Digital Markets Act, Data Act, and Data Governance Act. There is great scepticism about the effectiveness of a comprehensive AI regulation in Brazil without other supporting regulations. Moreover, critics emphasise that a comprehensive approach to AI regulation based on risk assessment and mitigation may be disproportionately restrictive, especially considering that Brazil is still lagging behind in terms of AI research and development. Therefore, a more pro-innovation approach might be a better fit. Nonetheless, Bill nº 2.338/2023 also encompasses other regulatory instruments that cannot be ignored: - It grants a number of individual rights to all those impacted by AI systems in the country, including a right to be informed of any interactions with those systems, a right to have any decisions made or supported by AI explained, and a right to contest AI-based decisions. - It creates a specific rule for civil liability for any damages caused by AI systems. - It stipulates rules and proceedings regarding the creation of regulatory sandboxes for AI in Brazil. ## Sources - PL da Inteligência Artificial (Bill 2338/2023), 10 December 2024: https://www25.senado.leg.br/web/atividade/materias/-/materia/157233 - Plano Brasileiro de Inteligência Artificial (PBIA) 2024-2028, 7 August 2024: https://www.gov.br/lncc/pt-br/assuntos/noticias/ultimas-noticias-1/plano-brasileiro-de-inteligencia-artificial-pbia-2024-2028 - Final Report of the Commission of Jurists to Support the Drafting of a Substitute Bill on Artificial Intelligence in Brazil, December 2022: https://www.stj.jus.br/sites/portalp/SiteAssets/documentos/noticias/Relato%CC%81rio%20final%20CJSUBIA.pdf - Estratégia Brasileira de Inteligência Artificial – EBIA, July 2021: https://www.gov.br/mcti/pt-br/acompanhe-o-mcti/transformacaodigital/arquivosinteligenciaartificial/ebia-documento_referencia_4-979_2021.pdf - Bill on the Use of AI in Brazil, 4 February 2020: https://www.congressonacional.leg.br/materias/materias-bicamerais/-/ver/pl-21-2020#:~:text=Ementa%3A,Brasil%3B%20e%20d%C3%A1%20outras%20provid%C3%AAncias. --- # AI Policy: Colombia (CO) Canonical page: https://aipolicy.ius.uzh.ch/compare/colombia Last updated: 2025-06-25T12:42:12.857Z ## General Description Colombia launched its first national AI policy in 2019, becoming one of the early adopters of an official governmental approach to artificial intelligence in Latin America. This policy emphasized evidence-based regulation and prioritized regulatory experimentation. It led to various capacity-building initiatives to support the development of experimental regulatory spaces. In parallel, Colombia also published an ethical framework establishing a set of principles for the use and deployment of AI in the public sector. In 2024, the government published a new national AI policy that renewed its focus on ethical guidelines for AI use and laid the groundwork for the development of a national AI law. As part of this effort, Colombia is now advancing toward the adoption of a national AI law through the Proyecto de Ley por medio del cual se regula la inteligencia artificial en Colombia para garantizar su desarrollo ético y responsable, introduced in May 2025 by the Ministry of Science, Technology and Innovation and the Ministry of ICT. This unified bill consolidates various earlier proposals and aims to provide a comprehensive legal framework for the ethical and sustainable development of AI. Drawing on the EU AI Act for inspiration, the Colombian bill emphasizes national priorities, such as human rights protections, equitable access to technology, and regulatory innovation. In this context, the use of AI regulatory sandboxes continues, now framed within risk-based regulatory testing and evidence gathering. The draft legislation sets out guiding principles, individual rights, risk classification mechanisms, and sector-specific governance mandates. It promotes transparency, accountability, and responsible innovation while introducing adaptive tools such as regulatory sandboxes. The bill is currently under legislative review and awaits debate and approval by the Colombian Congress. ## Steps Taken - Ongoing: Colombia remains actively engaged in international forums (OECD, UNESCO, GPAI) and seeks alignment with global ethical and governance frameworks. - May 2025: The Ministry of Science, Technology and Innovation and the Ministry of ICT submitted the [unified AI Bill](https://minciencias.gov.co/sala_de_prensa/minciencias-lidera-el-proyecto-ley-que-busca-regular-el-desarrollo-etico-seguro-y) to the Colombian Senate. - 2024: A [new national AI Strategy document](https://colaboracion.dnp.gov.co/CDT/Conpes/Econ%C3%B3micos/4144.pdf) was published, emphasizing regulatory innovation, ethics, and the need for a unified legislative approach. - 2023–2024: Multiple legislative bills were introduced, addressing a wide range of AI-related issues, from ethics and labor impacts to data governance and copyright. - 2023: The “Hoja de Ruta para la Adopción Ética y Sostenible de la Inteligencia Artificial” was published, outlining a strategic vision rooted in human rights, sustainability, and technological sovereignty. - 2020: The government published its [AI Ethical Framework](https://www.usergioarboleda.edu.co/wp-content/uploads/2021/11/Marco-etico-para-la-inteligencia-artificial-en-Colombia-Maestria-en-Inteligencia-artificial.pdf) for the public sector, outlining guiding principles for responsible use. - 2019: Colombia launched its first [National AI Strategy](https://colaboracion.dnp.gov.co/CDT/Conpes/Econ%C3%B3micos/3975.pdf), emphasizing evidence-based policymaking and regulatory experimentation. ## Approaches Colombia’s AI bill adopts a risk-based classification model influenced by the EU AI Act, while being tailored to the country’s specific context. AI systems will be classified as follows: - Prohibited AI systems: These are entirely banned due to their potential to cause significant harm. - High-risk AI systems: Subject to strict requirements before they can be marketed or deployed. - Limited-risk AI systems: Must comply with specific transparency obligations. - Low-risk AI systems: Subject to minimal regulatory requirements. Pre-implementation risk assessments will be required, particularly for high-risk systems, which must also undergo additional oversight and algorithmic impact assessments. The bill explicitly prohibits AI systems that manipulate user behavior, exploit vulnerabilities, or threaten human rights—such as social scoring mechanisms or discriminatory surveillance technologies. The bill mandates the creation of regulatory sandboxes to enable safe experimentation with AI systems under government supervision. A competent authority will be designated to oversee AI regulation, while sector-specific regulators will coordinate implementation efforts. ## Content While influenced by international frameworks, Colombia’s AI bill addresses national development gaps. It includes: - Individual rights: Including the right to be informed about AI use, explanation of algorithmic decisions, and the right to contest them. - Civil liability regime: Clear accountability for harm caused by AI systems. - Public sector innovation: Promoting AI use in education, agriculture, health, and justice while ensuring ethical safeguards. - Data governance and cross-border considerations: Addressing data privacy, sovereignty, and ethical data use. - Support for research and industry: Incentives for AI research and capacity-building in both public and private sectors. - Environmental and social equity provisions: Aligning AI adoption with sustainability and social inclusion goals. ## Sources - National Council of Economic and Social Policy – National Artificial Intelligence Policy 14 February 2025: https://colaboracion.dnp.gov.co/CDT/Conpes/Econ%C3%B3micos/4144.pdf - Ministry of Science, Technology and Innovation – Roadmap for the Development and Application of Artificial Intelligence in Colombia, February 2024: https://minciencias.gov.co/sites/default/files/upload/noticias/hoja_de_ruta_adopcion_etica_y_sostenible_de_inteligencia_artificial_colombia_0.pdf - Discussion Document – Ethical Framework for Artificial Intelligence in Colombia, August 2020: https://www.usergioarboleda.edu.co/wp-content/uploads/2021/11/Marco-etico-para-la-inteligencia-artificial-en-Colombia-Maestria-en-Inteligencia-artificial.pdf - National Council of Economic and Social Policy – National Policy for Digital Transformation and Artificial Intelligence, 8 November 2019: https://colaboracion.dnp.gov.co/CDT/Conpes/Econ%C3%B3micos/3975.pdf --- # AI Policy: European Union (BE) Canonical page: https://aipolicy.ius.uzh.ch/compare/european-union Last updated: 2025-06-30T07:58:51.830Z ## General Description The [EU AI Act](https://eur-lex.europa.eu/eli/reg/2024/1689/oj) entered into force on 1 August 2024, marking the world’s first comprehensive regulatory framework for AI. The Act categorizes AI systems into four risk levels: unacceptable risk, high risk, limited risk, and minimal risk. AI systems classified as unacceptable risk, such as social scoring and manipulative AI, are banned, while high-risk AI systems must comply with transparency, data governance, and human oversight requirements. The regulation introduces conformity assessments and certification requirements for AI systems before they can enter the EU market. Providers and users of high-risk AI systems must adhere to harmonized standards and undergo regular audits. The [European AI Office](https://digital-strategy.ec.europa.eu/en/policies/ai-office), established under the AI Act, is responsible for enforcement and supervision, working alongside national authorities. The EU aims to address AI-related liability issues by amending the [Product Liability Directive](https://eur-lex.europa.eu/eli/dir/2024/2853/oj/eng) and by introducing a [AI Liability Directive](https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex:52022PC0496). ## Steps Taken - The proposal of the [AI Liability Directive](https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex:52022PC0496) was withdrawn by the Commission on 21 February 2025. - On 6 February 2025 the Commission released key guidelines, including regarding the [definition of AI systems](https://digital-strategy.ec.europa.eu/en/library/commission-publishes-guidelines-ai-system-definition-facilitate-first-ai-acts-rules-application) and the [prohibited practices](https://www.dirittobancario.it/wp-content/uploads/2025/02/C_2025_884_2_EN_ACT_part1_v2_OlFHGWKJRcz1XPvgNW2ZVuggMw_112366.pdf), to aid compliance. - The new [Product Liability Directive](https://eur-lex.europa.eu/eli/dir/2024/2853/oj/eng) went into force on 13 December 2024 and includes rules on software products including AI systems. - The EU AI Act entered into force on 1 August 2024. The application of its provisions is staggered over the following months and years: - On 2 February 2025 the prohibitions on unacceptable AI practices (e.g., social scoring) and AI literacy obligations took effect. - On 2 August 2025, several provisions will take effect: - the rules for setting up and designating national authorities and bodies to certify high-risk AI systems (Articles 28–32). - the rules on general purpose AI models (Articles 51–55). Containing specific obligations (e.g. transparanecy and risk-management) to providers of such systems. - the rules on the establishment on the [EU AI Office](https://digital-strategy.ec.europa.eu/en/policies/ai-office) (Articles 64–70). It will serve a critical function in executing the AI Act, especially in relation to general-purpose AI models. - the rules on enforcement powers, market surveillance, and cooperation across the EU (Articles 88–100 and 102–112), excluding Article 101 (fines for general purpose AI Providers) and Article 78 (rules on protecting confidential info during enforcement). - - On 2 August 2027, the full set of rules for high-risk AI systems embedded in regulated products will take force. Manufacturers of things like AI-powered medical scanners or smart machinery must comply by this date. The extra time accounts for the complexity of aligning AI rules with existing product safety laws. - Since not all provisions of the EU AI Act will come into force immediately, the [AI Pact](https://digital-strategy.ec.europa.eu/en/policies/ai-pact) was established. It supports organisations to plan ahead for the implementation of the measures laid down in the AI Act. ## Approaches The EU created a comprehensive legal structure that follows a risk-based approach. It includes enforceable rules on governance (e.g., the [EU AI Office](https://digital-strategy.ec.europa.eu/en/policies/ai-office)), and penalties (e.g., fines up to 6% of global turnover for major breaches). The framework is complemented by adjacent laws (e.g., [AI Liability Directive](https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex:52022PC0496) (not in force)) and voluntary initiatives (e.g., [AI Pact](https://digital-strategy.ec.europa.eu/en/policies/ai-pact)) that aims for a cohesive ecosystem. ## Content The AI Act applies a risk-based approach and contains provisions that aim to reduce the risks that may be caused by AI systems. To this end, the act contains different rules for four risk categories: Unacceptable risk, high risk, limited risk and minimal risk (implicitly). **Unacceptable risk** **AI systems** are prohibited. Unlawful are the placing on the market, the putting into service or the use of: - an AI system that uses subliminal, manipulative, or deceptive techniques that distort people's behaviour by impairing their ability to make informed decisions, causing them to take actions they otherwise would not have taken, and leading to significant harm. - an AI system that exploits vulnerabilities of individuals (such as their age, disability, or specific social or economic situation) in a way that distorts their behaviour and causes or is likely to cause significant harm. - an AI system that evaluates or scores people over time based on their social behaviour or personal characteristics, and the score results in: unfair treatment in unrelated social contexts; or treatment that is unjustified or disproportionate to their behaviour. - an AI system that predicts a person's risk of committing a crime based solely on profiling or personality traits, unless it is used to assist humans who are already relying on objective, verifiable facts linked to criminal activity. - an AI system that builds or expands facial recognition databases by indiscriminately scraping images from the internet or CCTV footage. - an AI system is used to detect emotions in workplaces or educational institutions, unless it is for medical or safety reasons. - an AI system that categorises people based on biometric data (like facial features) to infer sensitive information such as race, political views, religious beliefs, or sexual orientation. (Exception: Labelling or filtering of lawfully acquired datasets in law enforcement). - an AI system that is used in law inforcement for ‘real-time’ remote biometric identification (like live facial recognition in public spaces), unless they are strictly necessary for: - Finding specific victims of serious crimes (e.g., human trafficking); - Preventing imminent and serious threats to life or terrorist attacks; or - Locating or identifying suspects of serious crimes punishable by at least four years in prison.. AI systems that negatively affect safety or fundamental rights are considered **high risk**. These systems fall into two categories: 1. AI systems that are used in products falling under the EU’s product safety legislation such as machinery, toys, aviation, cars, medical devices, lifts, and others; and 2. AI systems that stand for critical use cases and are considered to be high-risk, such as biometric and biometrics-based systems, management and operation of critical infrastructure, education and vocational training, employment, workers management and access to self-employment, access to and enjoyment of essential private services and public services and benefits, law enforcement, migration, asylum and border control management, and administration of justice and democratic processes. **High-risk AI systems** may only be placed on the market, put into service, or used if they comply with mandatory requirements. These requirements concern the quality and relevance of data sets, technical documentation and record-keeping, transparency and information for deployers, human oversight, as well as robustness, accuracy, and cybersecurity. To make sure, high risk AI systems meet these requirements, they must undergo a conformity assessment before being put on the market and such systems must be re-assessed throughout their lifecycle. **Limited risk AI systems** must comply with transparency requirements that allow users to make informed decisions. Namely, users should be made aware when they are interacting with an AI system. In addition to the transparency requirements applicable to limited risk AI systems, generative AI-systems should ensure transparency about the fact that the content is generated by an AI system and about the use of training data protected under copyright law. **Minimal risk AI** systems remain unregulated. ## Sources - Guidelines on the definition of an artificial intelligence system, 6 February 2025: https://digital-strategy.ec.europa.eu/en/library/commission-publishes-guidelines-ai-system-definition-facilitate-first-ai-acts-rules-application - Guidelines on prohibited artificial intelligence (AI) practices, 4 February 2025: https://www.dirittobancario.it/wp-content/uploads/2025/02/C_2025_884_2_EN_ACT_part1_v2_OlFHGWKJRcz1XPvgNW2ZVuggMw_112366.pdf - AI Pact, 25 September 2024: https://digital-strategy.ec.europa.eu/en/policies/ai-pact - AI Act, 13 June 2024: https://eur-lex.europa.eu/eli/reg/2024/1689/oj - Product Liability Directive, 23 October 2023: https://eur-lex.europa.eu/eli/dir/2024/2853/oj/eng - AI Liability Directive Proposal, 28 September 2022: https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex:52022PC0496 --- # AI Policy: Hong Kong (HK) Canonical page: https://aipolicy.ius.uzh.ch/compare/hong-kong Last updated: 2026-03-02T14:37:55.006Z ## General Description Hong Kong SAR has no standalone «AI Act». AI use is chiefly governed by the Personal Data (Privacy) Ordinance (PDPO) ([Cap. 486](https://www.elegislation.gov.hk/hk/cap528?utm_source=chatgpt.com)) and its six [Data Protection Principles](https://www.pcpd.org.hk/english/data_privacy_law/6_data_protection_principles/principles.html) (DPPs). AI-specific guidance is voluntary but anchored to the PDPO, primarily via the Office of the Privacy Commissioner for Personal Data (PCPD). ## Steps Taken There are several Foundational Policy and Cross-Sector-Guidance Documents: - The Office of the Privacy Commissioner for Personal Data (PCPD): - March 2025: [Checklist on Guidelines for the Use of Generative AI by Employees](https://www.pcpd.org.hk/english/news_events/media_statements/press_20250331.html). - June 2024: [Artificial Intelligence: Model Personal Data Protection Framework](https://www.pcpd.org.hk/english/resources_centre/publications/files/ai_protection_framework.pdf). - August 2021: [Guidance on the Ethical Development and Use of Artificial Intelligence](https://www.pcpd.org.hk/english/resources_centre/publications/files/guidance_ethical_e.pdf). - Digital Policy Office (DPO): - April 2025: [Hong Kong Generative Artificial Intelligence Technical and Application Guideline](https://www.digitalpolicy.gov.hk/en/our_work/data_governance/policies_standards/ethical_ai_framework/doc/HK_Generative_AI_Technical_and_Application_Guideline_en.pdf). - July 2024: [Ethical Artificial Intelligence Framework (Customised Version)](https://www.digitalpolicy.gov.hk/en/our_work/data_governance/policies_standards/ethical_ai_framework/). Furthermore, there are sector specific measures taken by the respective governmental bodies: - Financial Services: - November 2024: Securities and Futures Commission (SFC), [Circular to Licensed Corporations on Use of Generative Artificial Intelligence](https://apps.sfc.hk/edistributionWeb/gateway/EN/circular/intermediaries/supervision/doc?refNo=24EC55). - October 2024: Financial Services and the Treasury Bureau (FSTB), [Policy Statement on Responsible Application of Artificial Intelligence in the Financial Market](https://gia.info.gov.hk/general/202410/28/P2024102800154_475819_1_1730087238713.pdf). - August 2024: Hong Kong Monetary Authority (HKMA) launches [GenAI Sandbox](https://www.hkma.gov.hk/media/eng/doc/key-information/guidelines-and-circular/2024/20240920e1a1.pdf) to offer banks a risk-controlled environment to develop and test innovative solutions using AI. - 2019: Hong Kong Monetary Authority (HKMA), [High-Level Principles on Artificial Intelligence](https://brdr.hkma.gov.hk/eng/doc-ldg/docId/20191101-1-EN). - Healthcare: - January 2024: Department of Health, Medical Device Division (MDD), [Technical Reference TR-008: Artificial Intelligence Medical Devices](https://www.mdd.gov.hk/filemanager/common/mdacs/TR008.pdf). In Addition, there are some legislative developments: - [Copyright Ordinance (Cap. 528)](https://www.elegislation.gov.hk/hk/cap528) - Existing protection for “computer-generated” works; “author” is the person making necessary arrangements. - Proposed Text and Data Mining (TDM) exception following 2024 consultation; the [Legislative Council paper](https://www.ipd.gov.hk/en/copyright/current-topics/public-consultation-on-copyright-and-artificial/index.html) outlines: - Lawful access requirement. - Rights-holder opt-out (incl. machine-readable online reservations). - Disapplication where suitable licensing is available. Lastly, there are pro-innovation initiatives: - The Government has earmarked HK $ 1 billion and plans establishment of the [Hong Kong Artificial Intelligence Research and Development Institute (AIRDI)](https://www.hkgai.info/en) in 2026. ## Approaches Hong Kong’s AI governance landscape is anchored in a soft-law regulatory approach, with instruments issued by the Privacy Commissioner for Personal Data (PCPD), the Digital Policy Office (DPO), the Hong Kong Monetary Authority (HKMA), and the Securities and Futures Commission (SFC) designed to align with obligations under the Personal Data (Privacy) Ordinance (PDPO). Governance expectations are risk-based and principles-centric, emphasizing accountability, human oversight, transparency and interpretability, fairness, and data privacy and security. The Digital Policy Office (DPO) 2025 Guideline operationalizes this through a four-tier risk model: unacceptable, high, limited, and low. ## Content The [Personal Data (Privacy) Ordinance (PDPO) (Cap. 486)](https://www.elegislation.gov.hk/hk/cap486?utm_source=chatgpt.com) provides the binding baseline for AI deployments that involve personal data. It is implemented through the Data Protection Principles (DPPs): • DPP1: purpose and manner of collection, with data minimization that is directly relevant to model training and fine-tuning. • DPP2: accuracy and retention, which in practice translates into quality controls and defensible retention schedules for training corpora and logs. • DPP3: purpose limitation, meaning materially new training or re-use cases may require fresh notice and consent where the new purpose is not directly related. • DPP4: security safeguards for AI-related processing, typically requiring access controls, segregation, auditability, and secure handling of datasets, prompts, and outputs. • DPP5 (openness) and DPP6 (access and correction) also get pulled in to support transparency and data subject handling for AI-in-scope data and records. Within that foundation, Hong Kong relies heavily on soft-law instruments. [The Artificial Intelligence: Model Personal Data Protection Framework](https://www.pcpd.org.hk/english/resources_centre/publications/files/ai_protection_framework.pdf?utm_source=chatgpt.com) issued by the Office of the Privacy Commissioner for Personal Data (PCPD) structures organisational practice into four pillars: • Pillar 1: AI strategy and governance, including senior management approval, role clarity, internal policies, and staff training. • Pillar 2: risk assessment and human oversight, typically a privacy impact assessment-style review plus human-in-the-loop or human-on-the-loop controls calibrated to risk. • Pillar 3: customisation and lifecycle management, covering data governance, testing and validation, monitoring for drift and failure modes, and change management. • Pillar 4: stakeholder communication, including transparency measures, meaningful explanations where appropriate, and channels for challenge and human intervention. The [Hong Kong Generative Artificial Intelligence Technical and Application Guideline](https://www.digitalpolicy.gov.hk/en/our_work/data_governance/policies_standards/ethical_ai_framework/doc/HK_Generative_AI_Technical_and_Application_Guideline_en.pdf?utm_source=chatgpt.com) issued by the Digital Policy Office (DPO) (released April 2025 and subsequently updated) operationalises a four-tier risk model with corresponding control intensity: • Unacceptable: use cases to be prohibited. • High: rigorous governance, testing, and oversight, often associated with consequential domains such as advice or hiring. • Limited: controls centred on user-facing transparency and guardrails, for example chatbot disclosures and safe-use boundaries. • Low: lighter-touch controls such as self-certification and basic monitoring, for example commodity filtering use cases. A financial services example is the [Securities and Futures Commission (SFC) Circular – Use of generative AI language models](https://apps.sfc.hk/edistributionWeb/api/circular/list-content/circular/intermediaries/supervision/doc?lang=EN&refNo=24EC55&utm_source=chatgpt.com), which reinforces senior management accountability and competence, expects end-to-end AI model risk management (testing, validation, and ongoing review, with stronger human oversight for higher-risk use cases), elevates cybersecurity and data risk management (including resilience against adversarial manipulation and attention to training data quality and fairness), and stresses third-party risk governance (due diligence, contractual controls, and ongoing monitoring for vendors and outsourced model components). On intellectual property, the [Copyright Ordinance (Cap. 528)](https://www.elegislation.gov.hk/hk/cap528?utm_source=chatgpt.com) is relevant where outputs and training materials intersect with protected works, including the concept of protection for computer-generated works, where authorship is tied to the person who made the necessary arrangements for creation. Separately, Hong Kong’s planned text and data mining (TDM) exception is set out through [the Copyright and Artificial Intelligence Consultation Paper](https://www.ipd.gov.hk/filemanager/ipd/en/share/consultation-papers/Eng-Copyright-and-AI-Consultation-Paper-20240708.pdf?utm_source=chatgpt.com) and the Government’s proposed way forward in [LegCo Panel paper: Outcomes of Public Consultation and Proposed Way Forward](https://www.legco.gov.hk/yr2025/english/panels/ci/papers/ci20250218cb2-240-4-e.pdf?utm_source=chatgpt.com), with safeguards such as lawful access requirements, opt-out mechanisms for rights holders, and licensing carve-outs where they apply. ## Sources - Hong Kong (HKSAR) – Copyright Ordinance (Cap. 528), June 1997: https://www.elegislation.gov.hk/hk/cap528?utm_source=chatgpt.com - Hong Kong (HKSAR) – Personal Data (Privacy) Ordinance (Cap. 486), August 1995: https://www.elegislation.gov.hk/hk/cap486?utm_source=chatgpt.com - Financial Services and the Treasury Bureau (FSTB) – Policy Statement on Responsible Application of Artificial Intelligence in the Financial Market, October 2024: https://gia.info.gov.hk/general/202410/28/P2024102800154_475819_1_1730087238713.pdf?utm_source=chatgpt.com - Innovation, Technology and Industry Bureau (ITIB) – Hong Kong Innovation and Technology Development Blueprint, December 2022: https://www.gov.hk/en/about/abouthk/factsheets/docs/technology.pdf - Digital Policy Office (DPO) – Hong Kong Generative Artificial Intelligence Technical and Application Guideline, April 2025: https://www.digitalpolicy.gov.hk/en/our_work/data_governance/policies_standards/ethical_ai_framework/doc/HK_Generative_AI_Technical_and_Application_Guideline_en.pdf?utm_source=chatgpt.com - Office of the Privacy Commissioner for Personal Data (PCPD) – Checklist on Guidelines for the Use of Generative AI by Employees, March 2025: https://www.pcpd.org.hk/english/news_events/media_statements/press_20250331.html - Digital Policy Office (DPO) – Ethical Artificial Intelligence Framework (Customised Version), July 2024: https://www.digitalpolicy.gov.hk/en/our_work/data_governance/policies_standards/ethical_ai_framework/?utm_source=chatgpt.com - Office of the Privacy Commissioner for Personal Data (PCPD) – Artificial Intelligence: Model Personal Data Protection Framework, June 2024: https://www.pcpd.org.hk/english/resources_centre/publications/files/ai_protection_framework.pdf?utm_source=chatgpt.com - Office of the Privacy Commissioner for Personal Data (PCPD) – Guidance on the Ethical Development and Use of Artificial Intelligence, August 2021: https://www.pcpd.org.hk/english/resources_centre/publications/files/guidance_ethical_e.pdf - Hong Kong Monetary Authority (HKMA) – Generative Artificial Intelligence Sandbox – Practical Insights Report, 31 October 2025: https://brdr.hkma.gov.hk/eng/doc-ldg/docId/getPdf/20251031-5-EN/20251031-5-EN.pdf?utm_source=chatgpt.com - Hong Kong Monetary Authority (HKMA) – Generative Artificial Intelligence Sandbox, 16 July 2025: https://brdr.hkma.gov.hk/eng/doc-ldg/docId/20241118-1-EN?utm_source=chatgpt.com - Securities and Futures Commission (SFC) – Circular to Licensed Corporations on Use of Generative Artificial Intelligence, November 2024: https://apps.sfc.hk/edistributionWeb/gateway/EN/circular/intermediaries/supervision/doc?refNo=24EC55 - Hong Kong Monetary Authority (HKMA) – Circular on Generative Artificial Intelligence Sandbox, September 2024: https://www.hkma.gov.hk/media/eng/doc/key-information/guidelines-and-circular/2024/20240920e1a1.pdf?utm_source=chatgpt.com - Medical Device Division, Department of Health – Technical Reference TR-008: Artificial Intelligence Medical Devices, January 2024: https://www.mdd.gov.hk/filemanager/common/mdacs/TR008.pdf - Hong Kong Monetary Authority (HKMA) – High-level Principles on Artificial Intelligence, 1 November 2019: https://brdr.hkma.gov.hk/eng/doc-ldg/docId/getPdf/20191101-1-EN/20191101-1-EN.pdf - Legislative Council Panel on Commerce, Industry, Innovation and Technology – Discussion Paper on Consultation Outcomes and Proposed Way Forward, February 2025: https://www.legco.gov.hk/yr2025/english/panels/ci/papers/ci20250218cb2-240-4-e.pdf?utm_source=chatgpt.com - Commerce and Economic Development Bureau (CEDB)/Intellectual Property Department (IPD) – Public Consultation on Copyright and Artificial Intelligence: Consultation Paper, July 2024: https://www.ipd.gov.hk/en/copyright/current-topics/public-consultation-on-copyright-and-artificial/index.html?utm_source=chatgpt.com - Commerce and Economic Development Bureau (CEDB)/Intellectual Property Department (IPD) – Written Submissions received during the Public Consultation, July 2024: https://www.ipd.gov.hk/en/copyright/current-topics/public-consultation-on-copyright-and-artificial/index.html?utm_source=chatgpt.com --- # AI Policy: Israel (IL) Canonical page: https://aipolicy.ius.uzh.ch/compare/israel Last updated: 2026-02-10T13:47:06.693Z ## General Description Israel has not yet adopted comprehensive regulation in AI and does not have a dedicated "AI Act" similar to the EU. However, in recent years, the government has taken significant steps to develop a national policy on the subject, which reflects a structured approach that balances promoting innovation with regulatory [oversight](https://www.gov.il/BlobFolder/policy/ai_2023/en/Israels%20AI%20Policy%202023.pdf). Complementing these executive initiatives, Israel's Parliament (the Knesset) exercises ongoing oversight through its committee system, with the Subcommittee on Artificial Intelligence and Advanced Technologies serving as the primary forum for [parliamentary scrutiny](https://main.knesset.gov.il/en/news/pressreleases/pages/press26225e.aspx) of AI policy. This parliamentary oversight mechanism ensures democratic accountability and enables legislators to examine emerging challenges, review government strategies, and shape the national discourse on AI governance. The institutional framework for regulation has gradually evolved through several key government resolutions. In 2021, [under Government Resolution No. 212](https://www.gov.il/en/pages/ai_2023), the Ministry of Innovation, Science and Technology was tasked with leading Israel's national policy on AI, specifically in the areas of regulation, data governance, and ethics. In 2023, the government adopted ֵ[Resolution No.173](https://innovationisrael.org.il/wp-content/uploads/2023/10/Governmnet-Resoluion-No.-173.pdf) to strengthen Israel's technological leadership, which included measures to develop research, human capital, and infrastructure in the field of AI and to implement the use of AI tools in the civilian public sector. These efforts have gained international recognition: in the [2024 Global AI Index](https://www.tortoisemedia.com/data/global-ai#rankings), Israel ranked 9th overall and 2nd in relative performance, excelling in private investment and research and in the 2024 [Government AI Readiness Index](https://oxfordinsights.com/ai-readiness/ai-readiness-index/), it ranked 17th out of 174 countries, standing out in innovation and economic integration. In recent years, key policy documents have been published on the subject. In 2022, the Ministry of Innovation, Science and Technology, in collaboration with the Ministry of Justice, published a White Paper on AI, and in 2023, following public consultations, the first official policy on "[Artificial Intelligence Regulation and Ethics](https://www.gov.il/BlobFolder/policy/ai_23/he/2023%20Artificial%20Intelligence%20Regulation%20and%20Ethics%20Policy%20Principles%20Document.pdf)" was published. In parallel, several authorities have developed sector-specific policies or guidelines including the [financial sector](https://www.gov.il/en/pages/ai-financialsector), [intellectual property](https://www.gov.il/BlobFolder/legalinfo/machine-learning/he/18-12-2022.pdf), [privacy](https://www.gov.il/BlobFolder/rfp/ai_reg/he/ai_regu_draft.pdf), [education](https://meyda.education.gov.il/files/Pop/0files/lemida-metukshevet/AI/Guidelines.pdf), and [legal counsel](https://www.israelbar.biz/halishka_vehmechozot/lishkat_orche_din/mosdot_halishka/hamohza_harzit/545.aspx). A significant development occurred in September 2024, when Israel joined the [Council of Europe's AI Convention](https://www.coe.int/en/web/artificial-intelligence/the-framework-convention-on-artificial-intelligence). Israel's approach to regulation is characterized by the adoption of a "soft" model that favors ethical principles and voluntary standards over rigid legislation. The official policy explicitly states that "there is no room at this time for promoting regulation through horizontal legislation unique to the field of AI", preferring a principles-based, sector-specific approach, while maintaining the possibility of adopting horizontal legislation in the future if common challenges arise across sectors. ## Steps Taken As of May 2025, there is no AI act in Israel. However, the government has taken comprehensive steps towards developing AI policies through national frameworks, sector-specific guidelines, regulatory opinions, and proposed legislation. - Government Framework and National Strategy: - In February 2023, the Israeli government adopted [Resolution No. 173](https://innovationisrael.org.il/wp-content/uploads/2023/10/Governmnet-Resoluion-No.-173.pdf) to strengthen the technological leadership of the State of Israel. The resolution ordered the adoption of a series of steps to develop research, human capital, and infrastructure in the field of AI. - On March 15, 2022, [Amendment No. 130](https://main.knesset.gov.il/activity/legislation/laws/pages/lawbill.aspx?t=lawsuggestionssearch&lawitemid=2161538) to the Traffic Ordinance was published. The ordinance established an arrangement intended to lay a normative infrastructure for conducting trials in the operation of independent (autonomous) vehicles in commercial operation on Israeli roads. - On August 2021, [Government Decision No. 212](https://www.gov.il/en/pages/ai_2023) ("Decision 212") was adopted, mandating the Ministry of Innovation, Science and Technology (MIST) to lead the development of a national AI plan for Israel and to implement the first phase of the "TELEM Outline", a strategic initiative aimed at advancing AI research, human capital, infrastructure, and technological tools. The decision was further [expanded in 2023](https://www.gov.il/en/pages/ai_2023). - Policy Documents and Regulatory Approach: - In April 2024, Parliament's Research and Information Center submitted a [report](https://main.knesset.gov.il/EN/activity/mmm/National%20strategies%20and%20programs%20for%20artificial%20intelligence%20\(Summary\).pdf) analyzing AI strategies in Israel and leading nations, offering policy recommendations. - In December 2023, the Ministry of Innovation, Science and Technology issued a [professional document](https://www.gov.il/he/pages/ai_23) – ​​Principles of Policy, Regulation and Ethics in the Field of AI. The document reviews the legal and regulatory situation regarding AI in Israel and discusses the issues and challenges that arise in connection with AI. - In October 2022, the Minister of Innovation, Science and Technology published a policy document for public comment entitled “[Regulatory Policy and Ethics in the Field of Artificial Intelligence](https://www.gov.il/BlobFolder/news/most-news20223110/he/Regulatory%20and%20ethics%20policy%20document%20in%20the%20field%20of%20artificial%20intelligence%20in%20IsraelAI.pdf).” The document states that **there is no room at this stage for promoting regulation through horizontal legislation specific to the field of AI**. The document was mentioned in a position paper by the [Israeli Knesset Research and Information Center](https://fs.knesset.gov.il/globaldocs/MMM/11189446-7fc1-ed11-8159-005056aa4246/2_11189446-7fc1-ed11-8159-005056aa4246_11_20161.pdf). - - In July 2022, it was announced that Israel had formulated a "[National Plan for Artificial Intelligence](https://www.gov.il/he/pages/ai-data)" under the leadership of the Ministry of Innovation, Science and Technology and the Innovation Authority operating within the ministry. The aforementioned national plan was formulated following the work of the [AI and Data Science Committee](https://www.academy.ac.il/SystemFiles2015/14-3-21.pdf). The report was mentioned in a position paper by the [Israeli Knesset Research and Information Center](https://fs.knesset.gov.il/globaldocs/MMM/11189446-7fc1-ed11-8159-005056aa4246/2_11189446-7fc1-ed11-8159-005056aa4246_11_20161.pdf). The report opposes sector-specific "soft" regulatory tools like guidelines and standards. It emphasizes that financial entities remain fully responsible for AI systems, whether developed internally or externally. - In 2022, Israel’s Ministry of Innovation, Science and Technology, in collaboration with the Ministry of Justice (MOJ), published a [White Paper](https://www.gov.il/en/pages/most-news20221117) on AI; this was followed in 2023 by the release of MIST’s first policy on AI regulation and ethics, following public consultation. - International Commitments: - In September 2024, Israel signed the first international the [Council of Europe's AI Convention](https://www.coe.int/en/web/artificial-intelligence/the-framework-convention-on-artificial-intelligence), aimed at promoting responsible AI use and addressing risks to human rights, democracy, and the rule of law. - Sector-Specific Guidelines and Regulations: - [Education sector –](https://meyda.education.gov.il/files/Mazkirut_Pedagogit/Ivrit/instructionsai.pdf) In January 2025, The Ministry of Education's Innovation and Technology Administration has issued internal instructions for the use of AI in the education system. - [Banking Supervision sector](https://www.boi.org.il/media/mcyfqkkp/h2792.pdf) – In August 2024, the Supervisor of Banks issued a notice regarding risk management. The Supervisor of Banks announced that he is aware of the increasing use of AI models. This directive included guidelines regarding the implementation of AI models in banks, such as verification processes, benchmarking, robustness and stability testing (including research foundations), and analysis of model overrides. - [Israel Bar Association](https://www.israelbar.biz/halishka_vehmechozot/lishkat_orche_din/mosdot_halishka/hamohza_harzit/545.aspx) – In May 2024, the Israel Bar Association issued a statement on the use of AI. The statement stated that a lawyer who relied on the output of an AI platform while representing his client, without verifying the information or facts presented in this output, may be considered to have violated the duty of loyalty and devotion to which he owes, according to Section 54 of the Bar Association Law, and according to Rule 2 of the Bar Association Rules (Professional Ethics). - Privacy – October 2022: The Privacy Protection Authority (the Israeli privacy regulator) has published an [opinion](https://www.dataguidance.com/opinion/israel-ppas-new-opinion-regarding-disclosures-and) clarifying that the collection and use of personal data using algorithm or AI-based decision-making must conform with notification obligations. In April 2025, Israel's Privacy Protection Authority published a [draft guidance document](https://www.gov.il/BlobFolder/rfp/ai_reg/he/ai%20regu_draft.pdf) on applying Privacy Protection Law provisions to AI systems, which has been opened for public comment. The stated purpose of this guidance is to present the Privacy Protection Authority's interpretation of how provisions of the Privacy Protection Law apply to AI systems. - [IP sector](https://www.gov.il/BlobFolder/legalinfo/machine-learning/he/18-12-2022.pdf) – December 2022: The MOJ has published an opinion on IP challenges in the context of large scale AI models. In brief, the opinion concluded that apart from exceptional cases, the use of copyrighted materials for machine learning purposes as it applies to the learning process itself (rather than the output of machine-learning based systems) is generally permitted. - [Financial sector](https://www.gov.il/en/pages/ai-financialsector) – In July 2022, Israel's Ministry of Justice released a report examining AI in the financial sector. The report outlines a practical regulatory framework built around a three-tiered risk classification system. In the report AI systems are classified by risk level: low-risk systems face minimal requirements, medium-risk systems need enhanced governance and documentation, while high-risk systems require explainability, human oversight, regular reviews, and possible regulatory pre-approval. - Legislation Bills: - As of May 2025, there are two bills submitted to the Knesset (Israel’s parliament) dealing with specific issues relating to AI. It is significant to mention that these are private member bills, which have not yet been adopted as law. - [March 2025](https://www.law.co.il/media/computer-law/marking_of_ai_created_ads_5600_bill.pdf) – a bill proposing mandatory labeling of advertisements created or modified by AI. - [](https://www.law.co.il/media/computer-law/marking_of_ai_created_ads_5600_bill.pdf)[January 2025](https://www.law.co.il/media/computer-law/deep_fake_5677_private_bill.pdf) – the bill proposed establishing criminal penalties of 5–10 years for publishing AI-generated deepfakes aimed at sexually harming individuals, committing fraud, or influencing elections. - Parliamentary AI Oversight: - In July 2023, as part of its parliamentary oversight functions, Israel's Parliament established the Subcommittee on Artificial Intelligence and Advanced Technologies under its Science and Technology Committee. The subcommittee has examined AI's impact on labor markets, strategic investment needs, and national competitiveness. - [In February 2025](https://main.knesset.gov.il/en/news/pressreleases/pages/press26225e.aspx), think tanks, government agencies, and central bank representatives presented labor market forecasts, noting that while AI augments many professions, workforce displacement is inevitable. - [In January 2025](https://main.knesset.gov.il/en/news/pressreleases/pages/press15125q.aspx), it addressed the imperative for a national AI strategy and U.S. export restrictions on advanced chips, framing AI as essential to economic growth and national security. ## Approaches Israel's approach to AI regulation follows a [guideline first, legislation later](https://www.gov.il/en/pages/most-news20221117) policy, similar to practices in the US, Japan, UK, and Taiwan. This approach balances supporting new technology while having proper safety rules, allowing progress without too many restrictions as AI continues to develop quickly. The regulatory structure is founded on two complementary pillars: (1) national initiatives to develop research, human capital, and infrastructure in AI as outlined in Government Resolution No. 173 (February 2023); and (2) a principles-based regulatory framework emphasizing ethical guidelines and voluntary standards, with particular attention to high-risk applications. The current regulatory approach prioritizes sector-specific guidelines with flexible tools tailored to each industry's needs, while preserving the option for horizontal legislation if common challenges emerge across sectors. Several sector-specific frameworks have been developed, including a three-tiered risk classification system for the financial sector, age-appropriate guidelines for AI use in education, and professional ethics guidance for legal practitioners. These frameworks prioritize risk management, information disclosure, security measures, and data privacy protection. By grounding its approach in general ethical principles and soft regulatory tools, Israel seeks to ensure that its governance of AI remains responsive to evolving technologies, while upholding bias prevention, transparency, safety, accountability, and privacy. Israel has demonstrated international engagement by signing the first international AI Convention in September 2024, which promotes responsible AI use while addressing risks to human rights and democracy. The Ministry of Innovation serves as a central knowledge hub and coordination point, helping to prevent the regulatory fragmentation and oversight conflicts experienced by countries with dual-layered regulatory structures. ## Content Currently, Israel operates without an AI Act, instead working through various targeted regulatory initiatives and policy documents. Israel's policy focuses on a principles-based, sector-specific approach that relies on "soft" regulatory tools such as ethical guidelines and voluntary standards. This approach is based on the understanding that the field of AI should be allowed to develop and establish itself before imposing strict regulation, while emphasizing that this policy may change as the field evolves and its implications become clearer. In parallel with Israel's policy, we are also seeing formal legislative proposals emerge. In January 2025, the Knesset proposed criminal penalties for harmful AI deep-fakes, while in March 2025, another bill proposed mandatory AI labeling for advertisements. This development suggests we may see continued legislation regarding specific sectors in AI in the future. As AI applications expand across industries, Israel will likely introduce more targeted regulations to address emerging challenges in different sectors. ## Sources - Draft Guideline of the Privacy Protection Authority: Applicability of the Privacy Protection Law to Artificial Intelligence System, 28 April 2025: https://www.gov.il/BlobFolder/rfp/ai_reg/he/ai_regu_draft.pdf - Penal Law (Amendment – Deepfake Offenses) Bill, 5785–2025, 31 March 2025: https://www.law.co.il/media/computer-law/deep_fake_5677_private_bill.pdf - Bill for the Labeling of Advertisements Containing Content Created or Edited Using Artificial Intelligence (Amendments to Legislation), 5785–2025, 10 March 2025: https://www.law.co.il/media/computer-law/marking_of_ai_created_ads_5600_bill.pdf - Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law, 5 September 2024: https://www.coe.int/en/web/artificial-intelligence/the-framework-convention-on-artificial-intelligence - Supervisor of Banks – Notice Regarding Risk Management, 21 August 2024: https://www.boi.org.il/media/mcyfqkkp/h2792.pdf - Israeli Knesset Research and Information Center – Generative Artificial Intelligence: Opportunities, Risks, and Regulation, 7 June 2023: https://fs.knesset.gov.il/globaldocs/MMM/11189446-7fc1-ed11-8159-005056aa4246/2_11189446-7fc1-ed11-8159-005056aa4246_11_20161.pdf - Request for Information – The Inter-Agency Taskforce for Examining the Use of Artificial Intelligence and Machine Learning in the Financial Sector, 24 April 2023: https://www.gov.il/en/pages/ai-financialsector - Request for Information – The Inter-Agency Taskforce for Examining the Use of Artificial Intelligence and Machine Learning in the Financial Sector, 20 May 2023: https://www.gov.il/en/pages/ai-financialsector - Resolution No. 173 – Reinforcement of the technological leadership of the State of Israel, 24 February 2023: https://innovationisrael.org.il/wp-content/uploads/2023/10/Governmnet-Resoluion-No.-173.pdf - Israel's Policy on Artificial Intelligence – Regulation and Ethics 2023: https://www.gov.il/BlobFolder/policy/ai_2023/en/Israels%20AI%20Policy%202023.pdf - State of Israel Ministry of Justice – Opinion: Uses of Copyrighted Materials for Machine Learning, 18 December 2022: https://www.gov.il/BlobFolder/legalinfo/machine-learning/he/18-12-2022.pdf - Minister of Innovation, Science and Technology – Policy on Regulation and Ethics in the Field of Artificial Intelligence in Israel, 30 October 2022: https://www.gov.il/BlobFolder/news/most-news20223110/he/Regulatory%20and%20ethics%20policy%20document%20in%20the%20field%20of%20artificial%20intelligence%20in%20IsraelAI.pdf - Israel: PPA's new opinion regarding disclosures and transparency, 11 October 2022: https://www.dataguidance.com/opinion/israel-ppas-new-opinion-regarding-disclosures-and - Traffic Ordinance Amendment Law (No. 130), 15 March 2022: https://main.knesset.gov.il/activity/legislation/laws/pages/lawbill.aspx?t=lawsuggestionssearch&lawitemid=2161538 - The Artificial Intelligence and Data Science Committee – Report, March 2021: https://www.academy.ac.il/SystemFiles2015/14-3-21.pdf --- # AI Policy: Japan (JP) Canonical page: https://aipolicy.ius.uzh.ch/compare/japan Last updated: 2025-06-30T08:07:14.481Z ## General Description Currently, Japan is not moving toward comprehensive AI-specific regulation but rather seeks to manage AI-related risks by updating existing laws or clarifying their interpretation where necessary. The "[Interim Report](https://www8.cao.go.jp/cstp/ai/interim_report.pdf)" published by the government's AI Policy Study Group in February 2025 emphasizes respecting voluntary initiatives by businesses in AI risk management and leveraging existing sector-specific laws. Should new regulations become necessary, these would be implemented only to the extent truly required, remaining technologically neutral, avoiding hindrance to legitimate research, and minimizing burdens on businesses. Some describe this approach as Japan adopting a non-binding "soft-law approach," but that characterization is misleading. Although Japan's current policy does not involve enacting comprehensive AI-specific laws, existing regulations are relevant in numerous areas, such as privacy, copyright, as well as various safety regulations. In addition, legislative amendments anticipating AI use are progressing across sectors. The distinctive feature of Japan’s policy is less about restricting AI use and more about actively promoting AI research, development, and its practical application within regulated sectors. Additionally, in February 2025, the government submitted [a bill](https://www.meti.go.jp/english/press/2025/0207_001.html) to establish an AI Strategy Headquarters, chaired by the Prime Minister, aiming to strengthen strategic leadership functions across the entire government's AI policies. ## Steps Taken - Draft "[Guidelines for the Procurement and Utilization of Generative AI](https://public-comment.e-gov.go.jp/pcm/detail?CLASSNAME=PCMMSTDETAIL&Mode=0&id=290503241)" were published by the Digital Agency in March 2025. - "[Bill on Promoting Research, Development, and Utilization of AI-related Technologies](https://www.meti.go.jp/english/press/2025/0207_001.html)" was approved by the Cabinet and submitted to the Diet in February 2025. The bill aims to facilitate AI research, development, and use, primarily targeting government institutions. - "[Interim Report of the AI Policy Study Group was published by the Cabinet Office](https://www8.cao.go.jp/cstp/ai/interim_report.pdf)" in February 2025, emphasizing an approach that prioritizes sector-specific laws and voluntary initiatives by businesses rather than comprehensive AI-specific regulation. - "[Interim Report of the Study Group on Intellectual Property Rights in the AI Era](https://www.kantei.go.jp/jp/singi/titeki2/chitekizaisan2024/0528_ai.pdf)" was released by the Intellectual Property Strategy Headquarters in May 2024. - "[General Understanding on AI Policies](https://www8.cao.go.jp/cstp/ai/ai_senryaku/9kai/shiryo2-1.pdf)" was published by the Cabinet Office’s AI Strategy Team in May 2024. - "[AI Guidelines for Business](https://www.meti.go.jp/shingikai/mono_info_service/ai_shakai_jisso/pdf/20240419_9.pdf)" were jointly issued by the Ministry of Economy, Trade and Industry (METI) and the Ministry of Internal Affairs and Communications (MIC) in April 2024. These guidelines, which outline frameworks for AI risk management by businesses and government, were updated in March 2025. - "[General Understanding on AI and Copyright in Japan](https://www.bunka.go.jp/english/policy/copyright/pdf/94055801_01.pdf)" was published by the Agency for Cultural Affairs in March 2024. - As the chair of the G7, Japan led the Hiroshima AI Process in December 2023, reaching the [Hiroshima AI Process Comprehensive Policy Framework](https://www.soumu.go.jp/hiroshimaaiprocess/en/documents.html). G7 countries agreed on twelve "International Guiding Principles for All AI Actors" and eleven "International Codes of Conduct for Organizations Developing Advanced AI Systems." Currently, [more than 50](https://www.soumu.go.jp/hiroshimaaiprocess/en/supporters.html) countries and regions have been supporting the Hiroshima AI Process. - “[Social Principles of Human-Centric AI](https://www.cas.go.jp/jp/seisaku/jinkouchinou/pdf/humancentricai.pdf)” was issued by the Cabinet Office on 29 March 2019. ## Approaches Japan’s approach is risk-based, agile, and built on a multi-stakeholder process. Rather than establishing comprehensive AI-specific regulations, it prioritizes sector-specific responses by utilizing existing laws tailored to particular fields. The government respects businesses' voluntary AI governance initiatives and actively supports them through non-binding and flexible guidance. Concurrently, it places importance on collecting and disseminating best practices and information related to significant risks, aiming to ensure appropriate checks and balances across society. The 2025 interim report explicitly states that the introduction of new regulations should be limited to areas where risks cannot be effectively managed by businesses’ voluntary efforts alone. It stresses a case-by-case, risk-based approach, rather than uniformly applying regulation based on AI model scale metrics (e.g., floating-point operations per second (FLOPS) or user numbers). Japan actively contributes to international norm-setting initiatives, notably through the Hiroshima AI Process initiated during its 2023 G7 presidency. It respects principles agreed upon within various international frameworks, including the UN Global Digital Compact, the Council of Europe’s Framework Convention on AI, the OECD AI Principles, and the Global Partnership on AI (GPAI) resolutions. To ensure consistency and interoperability of international rules, Japan promotes active participation in international standardization activities by bodies such as the International Organization for Standardization (ISO) and the International Electrotechnical Commission (IEC). It also emphasizes the importance of integrating and sharing knowledge domestically and internationally through collaboration among AI safety institutes established in various countries. The Japanese government positions "agile governance"– a flexible, multi-stakeholder governance process – as a fundamental policy approach for the digital society and actively applies this principle in its AI governance efforts. ## Content Legal amendments are progressing in various fields with AI utilization in mind. Regarding autonomous driving, the amended Road Traffic Act was enacted in April 2023. This allows businesses with prefectural public safety commission approval to perform completely autonomous driving without human supervision or intervention (Level 4) under certain conditions. In finance, the Installment Sales Act was amended in 2020, allowing certified credit companies to use algorithms for credit assessments. The 2017 amendment to the Financial Instruments and Exchange Act requires businesses engaging in algorithmic high-frequency trading to register, conduct risk management processes, and produce and submit business reports. In legal services, AI use is advancing. Questions have been raised about whether AI-assisted contract drafting and review, known as "legal tech," constitutes unauthorized practice of law. The Ministry of Justice issued guidelines in August 2023 clarifying that the main functions of currently available contract creation and review services do not violate the Attorney Act. In healthcare, systems that assist doctors, such as AI for diagnosis, fall under "Software as a Medical Device (SaMD)" that require approval under the Pharmaceutical and Medical Device Act. This process can take over five years, but reforms to allow sales within a year of development are being discussed. Considering AI's continuous updates, a two-stage system is under consideration: initially allowing sales with minimal examination and later re-evaluating approval based on post-use data. Beyond these industry-specific regulatory changes and clarifications, there are cross-sectoral situations where AI substitution is expected. These involve what's termed "analog regulations," such as mandatory visual inspections, physical audits, regular checks, and onsite regulations. There are over 10,000 laws and notifications mandating analog compliance methods. Now, a collective amendment initiative was conducted by the government's Digital Adhoc Administrative Investigation Committee, resulting in the amended Digital Procedure Act in June 2023. In October of the same year, a technology map listing technologies that can replace these analog regulations was published. About copyright, Japan made a groundbreaking legal amendment in 2017. It stipulates that using copyrighted works for AI training by automatically downloading or processing data without human enjoyment of the work's expression, does not generally constitute copyright infringement (exceptions apply if it unfairly harms the copyright holder's interests, and copyright holders can prohibit such uses). However, this applies only to the AI training stage. Whether the output of generative AI infringing copyright when it resembles existing works is a topic of current debate. The key issue is interpretations of what it means for AI to "rely" on existing works. The "Interim Report of the AI Policy Study Group," published in February 2025, outlines specific directions for Japan's future legal framework. It emphasizes strengthening the government's strategic leadership functions and formulating strategic plans; ensuring transparency and appropriateness, including the introduction of information disclosure requirements; voluntary evaluation and AI governance certification for AI systems; investigation of major AI-related incidents and public disclosure of findings; promotion of AI adoption by the government, along with the development of procurement guidelines; and enhanced sector-specific measures by relevant ministries to manage risks related to life and physical safety, system stability, and national security. The "Bill on Promoting Research, Development, and Utilization of AI-related Technologies," approved by the Cabinet in February 2025, establishes an AI Strategy Headquarters chaired by the Prime Minister and mandates the formulation of a national AI basic plan. Furthermore, the bill authorizes the government to collect information on AI research, development, and associated risks, and to use this information as the basis for providing guidance, advice, information sharing, and other necessary measures aimed at research institutions, AI service providers, and other stakeholders. In March 2025, the Digital Agency published draft "Guidelines for the Procurement and Utilization of Generative AI," outlining criteria for determining high-risk usage within government contexts, along with checklists for risk management and contractual considerations. ## Sources - Guidelines for the Procurement and Utilization of Generative AI, March 2025: https://public-comment.e-gov.go.jp/pcm/detail?CLASSNAME=PCMMSTDETAIL&Mode=0&id=290503241 - Bill on Promoting Research, Development, and Utilization of AI-related Technologies, 7 February 2025: https://www.meti.go.jp/english/press/2025/0207_001.html - Social Principles of Human-Centric AI, 29 March 2019: https://www.cas.go.jp/jp/seisaku/jinkouchinou/pdf/humancentricai.pdf - Council for AI Strategy/Study Group on the Legal and Institutional Framework for AI – Interim Report, 4 February 2025: https://www8.cao.go.jp/cstp/ai/interim_report.pdf - General Understanding on AI and Copyright in Japan – Overview, May 2024: https://www.bunka.go.jp/english/policy/copyright/pdf/94055801_01.pdf - Study Group on Intellectual Property Rights in the AI Era – Interim Report, May 2024: https://www.kantei.go.jp/jp/singi/titeki2/chitekizaisan2024/0528_ai.pdf - General Understanding on AI Policies, May 2024: https://www8.cao.go.jp/cstp/ai/ai_senryaku/9kai/shiryo2-1.pdf - AI Guidelines for Business, 19 April 2024: https://www.meti.go.jp/shingikai/mono_info_service/ai_shakai_jisso/pdf/20240419_9.pdf - Hiroshima AI Process Comprehensive Policy Framework, December 2023: https://www.soumu.go.jp/hiroshimaaiprocess/en/documents.html --- # AI Policy: Nigeria (NG) Canonical page: https://aipolicy.ius.uzh.ch/compare/nigeria Last updated: 2025-06-30T08:08:32.507Z ## General Description Nigeria is actively navigating the complexities of artificial intelligence (AI) governance, although it has yet to enact a specific, overarching AI law. The country's approach is characterized by an evolving legislative landscape, with significant momentum in both the executive and legislative branches of government. Over the past few years, several AI-related bills have been introduced in the National Assembly, reflecting the growing [demand](https://guardian.ng/news/10th-nass-to-provide-legal-framework-for-ai-regulation/) from the legislature for a dedicated regulatory framework for AI. In a significant move toward creating a unified legislative front, four separate AI bills were [consolidated](https://nass.gov.ng/documents/download/11167) in the House of Representatives in December 2024. These were: the National AI and Robotics Sciences Bill (2023), the Control of Usage of Artificial Intelligence Technology Bill (2023), the Artificial Intelligence and Robotics Research Regulatory Agency Bill (2021), and the National Artificial Intelligence Regulatory Authority Bill (2024). The legislative push continued into 2025 with the introduction of two more pivotal bills in the Senate: the [National Artificial Intelligence Commission (Establishment) Bill](https://legis360.org/app/bill?id=571), which proposes the creation of a national body to oversee AI, and the [Nigerian Digital Sovereignty and Fair Data Compensation Bill](https://legis360.org/app/bill?id=564). This latter bill notably seeks to mandate that foreign companies using Nigerian data for AI training contribute 2% of their local annual revenue to a national AI development fund and conduct at least 30% of related research and development within Nigeria. Complementing these legislative efforts, the executive branch has released a comprehensive draft [National AI Strategy](https://ncair.nitda.gov.ng/wp-content/uploads/2024/08/National-AI-Strategy_01082024-copy.pdf), signaling a clear, top-down direction for harnessing AI for socio-economic development while establishing guardrails for its ethical implementation. ## Steps Taken Nigeria has adopted a multi-pronged, proactive approach to AI governance, engaging in the development of national strategies, international diplomacy, and sector-specific rulemaking: - In 2024, the government published a draft [National AI Strategy](https://ncair.nitda.gov.ng/wp-content/uploads/2024/08/National-AI-Strategy_01082024-copy.pdf) that outlines a vision for responsible and inclusive innovation. Key pillars of the strategy's governance plan include establishing a formal regulatory body, adopting national AI ethics principles, developing a robust policy framework, and implementing a comprehensive risk management system for the deployment of AI. The final version is expected to be published in 2025. - In November 2020, the [National Centre for Artificial Intelligence and Robotics](https://ncair.nitda.gov.ng/) was established to champion research and development in AI and other emerging technologies. - Nigeria is an active participant in global AI dialogues. It is a signatory to the [Bletchley Declaration](https://www.gov.uk/government/publications/ai-safety-summit-2023-the-bletchley-declaration/the-bletchley-declaration-by-countries-attending-the-ai-safety-summit-1-2-november-2023), the [African Declaration on AI](https://c4ir.rw/docs/Africa-Declaration-on-Artificial-Intelligence.pdf), and the Statement on [Inclusive and Sustainable AI for People and the Planet](https://altadvisory.africa/2025/02/13/six-african-states-global-statement-on-inclusive-and-sustainable-ai/). The country has also endorsed the [United Kingdom’s guidelines for secure AI development](https://www.ncsc.gov.uk/collection/guidelines-secure-ai-system-development) and signed the [Charter of Paris for the General Interest of Artificial Intelligence](https://www.elysee.fr/en/emmanuel-macron/2025/02/11/the-paris-charter-on-artificial-intelligence-in-the-public-interest). There have also been sector-specific and domain interventions: - In May 2025, the National Human Rights Commission [announced](https://thefact.ng/nhrc-other-push-for-ethical-ai-governance-in-nigeria/) that it is monitoring the human rights impacts of AI and will establish a dedicated unit, preparing to mandate human rights due diligence for tech companies. - In May 2025, the Central Bank of Nigeria [published]() the Draft Baseline Standards for Automated AML Solutions 2025. - In March 2025, the Nigerian Data Protection Commission [incorporated](https://ndpc.gov.ng/wp-content/uploads/2025/03/NDP-ACT-GAID-2025-MARCH-20TH.pdf) emerging technologies into its General Application and Implementation Directive, requiring data controllers and processors to embed data protection considerations into the development, deployment, and use of their services. - In March 2024, the Federal Competition and Consumer Protection Commission (FCCPC) [emphasised](https://www.thisdaylive.com/2024/03/19/fccpc-unregulated-ai-may-exploit-consumers-pose-cyber-security-risks/) the importance of regulating AI to safeguard consumers from unintended consequences, with a focus on promoting fair and responsible AI development in emerging economies. - In 2023, the Securities and Exchange Commission [published](https://sec.gov.ng/wp-content/uploads/2023/04/Rules-on-Robo-Advisory-Services_Executed-30-August-2021.pdf) rules governing the use of robo-advisory services in the financial sector. At the sub-national level, the Lagos State government has [announced](https://www.ecofinagency.com/news/1905-46871-lagos-to-release-nigeria-s-first-guidelines-for-artificial-intelligence) plans to introduce its guidelines for AI. ## Approaches Nigeria appears to be pursuing a hybrid regulatory approach. The significant and ongoing legislative activity suggests that the country will likely enact a foundational, AI-specific law to establish a clear governance structure and a dedicated regulatory authority. However, this horizontal legislation is expected to operate in concert with a vertical, sector-specific enforcement model. Existing regulators in finance, data protection, human rights, and consumer protection are already extending their mandates to address the risks and opportunities of AI within their domains. This approach allows for both a consistent, nationwide set of principles and the nuanced application of rules tailored to the specific contexts of different industries. It also anticipates that existing laws in areas like cybersecurity, competition, and product safety will be amended and strengthened to address emerging AI-related challenges. ## Content As Nigeria has not yet passed a specific law on AI, there is currently no legislative content to analyze. The regulatory landscape is defined by the draft National AI Strategy and the various bills before the National Assembly, which collectively indicate the future direction of AI governance in the country. ## Sources - SB731 – National Artificial Intelligence Commission (Establishment) Bill, 2025: https://legis360.org/app/bill?id=571 - SB722 – Nigerian Digital Sovereignty and Fair Data Compensation Bill, 2025: https://legis360.org/app/bill?id=564 - Central Bank of Nigeria – Exposure of Draft Baseline Standards for Automated Anti-Money Laundering (AML) Solutions: Request for Comments, 20 May 2025: https://www.cbn.gov.ng/Out/2025/CCD/Exposure Draft on Baseline Standards for Automated AML Solutions.pdf - The Africa Declaration on Artificial Intelligence, 4 April 2025: https://c4ir.rw/docs/Africa-Declaration-on-Artificial-Intelligence.pdf - Statement on Inclusive and Sustainable AI, 13 February 2025: https://altadvisory.africa/2025/02/13/six-african-states-global-statement-on-inclusive-and-sustainable-ai/ - The Paris Charter on Artificial Intelligence in the Public Interest, 11 February 2025: https://www.elysee.fr/en/emmanuel-macron/2025/02/11/the-paris-charter-on-artificial-intelligence-in-the-public-interest - NDPC – Nigeria Data Protection Act (NDP Act) 2023 – General Application and Implementation Directive (GAID), 2025: https://ndpc.gov.ng/wp-content/uploads/2025/03/NDP-ACT-GAID-2025-MARCH-20TH.pdf - House of Representatives – Federal Republic of Nigeria – Order Paper – Consolidated AI Bills, 3 December 2024: https://nass.gov.ng/documents/download/11167 - NAIS – National Artificial Intelligence Strategy, August 2024: https://ncair.nitda.gov.ng/wp-content/uploads/2024/08/National-AI-Strategy_01082024-copy.pdf - Bletchley Declaration, 2 November 2023: https://www.gov.uk/government/publications/ai-safety-summit-2023-the-bletchley-declaration/the-bletchley-declaration-by-countries-attending-the-ai-safety-summit-1-2-november-2023 - SEC Nigeria – Rule on Robo-Advisory Services, 2023: https://sec.gov.ng/wp-content/uploads/2023/04/Rules-on-Robo-Advisory-Services_Executed-30-August-2021.pdf --- # AI Policy: Rwanda (RW) Canonical page: https://aipolicy.ius.uzh.ch/compare/rwanda Last updated: 2025-06-26T12:46:40.070Z ## General Description Rwanda has positioned itself as a forward-thinking country on the African continent, strategically embracing artificial intelligence (AI) as a core component of its national development agenda. While the country does not currently have a standalone AI law, it is taking deliberate steps to build a robust governance framework. A key legislative development is the ongoing process to [amend](https://www.newtimes.co.rw/article/25203/news/technology/rwanda-set-to-integrate-ai-into-ict-legal-framework) the 2016 Information and Communication Technology (ICT) law to integrate specific considerations for AI and other emerging technologies. This legislative reform is guided by a clear executive vision, articulated in the country's comprehensive [National AI Policy](https://www.minict.gov.rw/index.php?eID=dumpFile&t=f&f=67550&token=6195a53203e197efa47592f40ff4aaf24579640e). This policy document serves as a roadmap, outlining an ambitious plan to leverage AI for economic growth and improved quality of life. To further operationalize its commitment to responsible innovation, the government also plans to [publish](https://www.newtimes.co.rw/article/17921/news/technology/how-rwanda-is-regulating-artificial-intelligence) Guidelines on the Ethical Development and Implementation of AI, ensuring that development aligns with societal values. ## Steps Taken Rwanda's actions demonstrate a structured and coherent effort to build an AI ecosystem founded on strong policy, ethical principles, and effective institutions: - In May 2025, the government published a [Data Sharing Policy](https://www.minict.gov.rw/index.php?eID=dumpFile&t=f&f=130290&token=2f207763665aa480e65386c025175bf4330b9a5c) to facilitate the secure and efficient exchange of data between government institutions, creating a critical resource for AI development. - In September 2024, Rwanda [collaborated](https://www.minict.gov.rw/index.php?eID=dumpFile&t=f&f=107417&token=6fe9f09b32c074900f09dd3f1095a8eed2551075) with Singapore to develop and launch the AI Governance Playbook for Small States. This practical guide is designed to help small nations address their unique challenges in adopting and governing AI, promoting a trusted and collaborative ecosystem. - The government has [established](https://coingeek.com/rwanda-pushes-for-safe-and-responsible-ai-development/) a Responsible AI Office. This central body is mandated to coordinate the implementation of the National AI Policy, championing the responsible and effective use of AI across the country. - In 2023, Rwanda [published](https://www.minict.gov.rw/index.php?eID=dumpFile&t=f&f=67550&token=6195a53203e197efa47592f40ff4aaf24579640e) its National AI Policy, outlining its vision to become a global innovator in responsible and inclusive AI. The policy emphasizes building human capital, developing robust data and computing infrastructure, and driving AI adoption in both the public and private sectors. The National AI Policy places a strong emphasis on ethics as a prerequisite for ensuring that the benefits of AI are distributed equitably. This focus will be further detailed in the forthcoming ethical guidelines. Rwanda is an engaged member of the global community on AI governance. It is a signatory to the [Bletchley Declaration](https://www.gov.uk/government/publications/ai-safety-summit-2023-the-bletchley-declaration/the-bletchley-declaration-by-countries-attending-the-ai-safety-summit-1-2-november-2023), the [African Declaration on AI](https://c4ir.rw/docs/Africa-Declaration-on-Artificial-Intelligence.pdf), and the [Statement on Inclusive and Sustainable AI for People and the Planet](https://altadvisory.africa/2025/02/13/six-african-states-global-statement-on-inclusive-and-sustainable-ai/). ## Approaches Rwanda is adopting a foundational and adaptive approach to AI regulation. Instead of immediately creating a new, separate AI law from scratch, the government's primary strategy is to modernize its existing legal frameworks. The amendment of the 2016 ICT law is central to this approach, allowing the country to build upon a solid legal foundation while adapting it to the unique challenges of AI. This strategy will be complemented by the expansion of mandates for various existing regulatory agencies, empowering them to oversee AI within their respective sectors. This indicates a preference for a co-regulatory model, where broad legal principles are established in foundational laws, such as the ICT Law. This approach enables both stability and the flexibility necessary to respond to the rapid evolution of AI technology. ## Content As Rwanda has not yet passed a specific law on artificial intelligence, there is currently no legislative content to analyze. The country's regulatory direction is primarily shaped by its National AI Policy and the announced intention to formally incorporate AI and other emerging technologies into the existing ICT law. ## Sources - Ministry of ICT and Innovation – The National Data Sharing Policy, May 2025: https://www.minict.gov.rw/index.php?eID=dumpFile&t=f&f=130290&token=2f207763665aa480e65386c025175bf4330b9a5c - The Africa Declaration on Artificial Intelligence, 4 April 2025: https://c4ir.rw/docs/Africa-Declaration-on-Artificial-Intelligence.pdf - Statement on Inclusive and Sustainable AI, 13 February 2025: https://altadvisory.africa/2025/02/13/six-african-states-global-statement-on-inclusive-and-sustainable-ai/ - AI Playbook for Small States – 22 September 2024: https://www.minict.gov.rw/index.php?eID=dumpFile&t=f&f=107417&token=6fe9f09b32c074900f09dd3f1095a8eed2551075 - Bletchley Declaration, 2 November 2023: https://www.gov.uk/government/publications/ai-safety-summit-2023-the-bletchley-declaration/the-bletchley-declaration-by-countries-attending-the-ai-safety-summit-1-2-november-2023 - Ministry of ICT and Innovation – The National AI Policy, 2022: https://www.minict.gov.rw/index.php?eID=dumpFile&t=f&f=67550&token=6195a53203e197efa47592f40ff4aaf24579640e --- # AI Policy: South Africa (ZA) Canonical page: https://aipolicy.ius.uzh.ch/compare/south-africa Last updated: 2026-04-28T14:45:05.801Z ## General Description South Africa is developing AI governance through a layered framework rather than a single statute. As a constitutional democracy, its approach is anchored in the 1996 Constitution, supported by existing data protection and cyber laws, and supplemented by emerging AI-specific policy instruments. As an upper-middle-income country, South Africa combines a relatively sophisticated legal and financial system with deep inequality and uneven state capacity. These conditions produce a policy approach that is both ambitious and pragmatic. South Africa has played an active role in regional and multilateral AI initiatives. It led the African Union’s Smart Africa AI Blueprint during its AU chairship in 2020 and contributed to the [AU Continental AI Strategy](https://au.int/en/documents/20240809/continental-artificial-intelligence-strategy) in 2024. A defining feature of its approach is the integration of AI regulation with digital public infrastructure (DPI), which includes digital identity, data exchange and digital payments. DPI is treated as the foundation for public sector AI applications. This reflects a development-oriented framing of AI as a tool to improve service delivery, reduce fraud and expand inclusion, subject to constitutional safeguards. ## Steps Taken South Africa’s AI and adjacent policy framework has developed incrementally rather than through a single legislative act: - 2026: Publication of the [Draft National AI Policy](https://www.gov.za/sites/default/files/gcis_document/202604/54477gen3880.pdf) for public comment. This substantially expands the 2024 Framework. However, the draft policy was retracted after it emerged that the policy contains hallucinated AI-generated references. - 2025: Release of the [Digital Transformation Roadmap](https://www.gov.za/sites/default/files/gcis_document/202505/south-africas-roadmap-digital-transformation-government.pdf), operationalising four DPI components as the backbone for AI-enabled public services. - 2024: Publication of the National AI Policy Framework, structured around nine thematic pillars. - 2024: Launch of the South African National AI Plan, a planning and discussion document. - 2020: Leadership of the [Smart Africa AI Blueprint](https://smartafrica.org/knowledge/artificial-intelligence-for-africa/) during South Africa’s AU chairship. - 2020: Publication of the Presidential Commission on the [Fourth Industrial Revolution report](https://www.gov.za/sites/default/files/gcis_document/202010/43834gen591.pdf), placing AI on the national policy agenda. - 2013: Enactment of the [Protection of Personal Information Act (POPIA)](https://www.gov.za/sites/default/files/gcis_document/201409/3706726-11act4of2013protectionofpersonalinforcorrect.pdf), which gives effect to the constitutional right to privacy and regulates the processing of personal data, including automated processing. It remains the primary statute applicable to AI systems handling personal information. POPIA came into full effect in 2020. ## Approaches **Constitutional priority** As a constitutional democracy, South Africa’s AI governance is directed by the [Constitution](https://www.justice.gov.za/constitution/SAConstitution-web-eng.pdf), and in particularly by the Bill of Rights. Key provisions include: - Section 9 (equality), addressing bias and discrimination - Section 10 (human dignity), supporting human-centred AI - Section 12 (freedom and security), relevant to biometric data use - Section 14(d) (privacy), relevant to consent and data minimisation - Section 16(1) (freedom of expression), relevant to content moderation - Section 32(1) (access to information), supporting transparency and explainability **Layered and sectoral model** South Africa is moving towards a system implemented through sector-specific regulators rather than a single horizontal AI law, as outlined in the retracted [Draft National AI Policy](https://www.gov.za/sites/default/files/gcis_document/202604/54477gen3880.pdf). **Institutionally distributed governance** Responsibility is distributed across multiple departments, including Communications and Digital Technologies, Public Service and Administration, Justice, Home Affairs and the National Planning Commission. These are supported by entities such as State Information Technology Agency (SITA), Government Information Technology Officers Council (GITOC), the Government Chief Information Officer and the National Advisory Council on Innovation (NACI). ## Content The core of South Africa’s emerging AI regime was supposed to be set out in the [Draft National Artificial Intelligence Policy](https://www.gov.za/sites/default/files/gcis_document/202604/54477gen3880.pdf), approved by Cabinet and published for a 60-day public comment period on 10 April 2026. However, due to hallucinated AI references in the draft policy it was retracted on shortly after its release. While an updated policy is under consultation, existing legislation continues to carry most regulatory weight. The most significant is Protection of Personal Information Act ([POPIA](https://www.gov.za/sites/default/files/gcis_document/201409/3706726-11act4of2013protectionofpersonalinforcorrect.pdf)), which gives effect to the constitutional right to privacy and remains the central statute for AI systems processing personal data. It sets conditions for lawful processing, including data minimisation and purpose limitation, and provides enhanced protection for children. Section 71 is particularly relevant, according to which a subject may not be subject to a decision that results in legal consequences or affects the person to a substantial degree, where that decision is based solely on automated processing intended to profile the person. POPIA is enforced by the Information Regulator and is likely to be the primary vehicle for early AI enforcement. However, its provisions on automated decision-making and transparency are less granular than comparable regimes such as the EU or UK GDPR, a gap the Draft Policy does not yet fully resolve. A broader legislative framework continues to apply to AI systems, including: - the [Cybercrimes Act 19 of 2020](https://www.gov.za/sites/default/files/gcis_document/202106/44651gon324.pdf), addressing cyber fraud, forgery and harmful data messages - the [Electronic Communications and Transactions Act 25 of 2002](https://www.gov.za/sites/default/files/gcis_document/201409/a25-02.pdf), governing electronic transactions and intermediary liability - the [Consumer Protection Act 68 of 2008](https://www.gov.za/sites/default/files/32186_467.pdf), relevant to automated decision-making and pricing - the [Copyright Act 98 of 1978](https://www.gov.za/sites/default/files/gcis_document/201504/act-98-1978.pdf) and the pending Copyright Amendment Bill, addressing authorship and data use in training models - the [Competition Act 89 of 1998](https://www.gov.za/sites/default/files/gcis_document/201409/a89-98.pdf), including risks of algorithmic collusion and data concentration - the [Promotion of Access to Information Act 2 of 2000 and Promotion of Administrative Justice Act 3 of 2000](https://www.gov.za/sites/default/files/gcis_document/201409/a2-000.pdf), which provide mechanisms for transparency and review of algorithmically assisted decisions - the [National Cybersecurity Policy Framework](https://www.gov.za/sites/default/files/gcis_document/201512/39475gon609.pdf) (2015) and [National Data and Cloud Policy](https://www.gov.za/sites/default/files/gcis_document/202406/50741gen2533.pdf) (2024), addressing data governance and resilience ## Sources - Department of Communications and Digital Technologies, Draft National Artificial Intelligence Policy (published for public comment,), 10. April 2026: https://www.gov.za/sites/default/files/gcis_document/202604/54477gen3880.pdf - The Presidency, Republic of South Africa, MyMzansi Digital Government Roadmap, 2025: https://www.gov.za/sites/default/files/gcis_document/202505/south-africas-roadmap-digital-transformation-government.pdf - Department of Communications and Digital Technologies, South Africa National Artificial Intelligence Policy Framework, 25. October 2024: https://www.dcdt.gov.za/sa-national-ai-policy-framework/file/338-sa-national-ai-policy-framework.html - Presidential Commission on the Fourth Industrial Revolution, Report (Government Gazette No. 43834, 2020), 23. October 2020: https://www.gov.za/sites/default/files/gcis_document/202010/43834gen591.pdf - Republic of South Africa, Protection of Personal Information Act 4 of 2013 (POPIA): https://www.gov.za/sites/default/files/gcis_document/201409/3706726-11act4of2013protectionofpersonalinforcorrect.pdf - African Union, Continental Artificial Intelligence Strategy, 9. August 2024: https://au.int/en/documents/20240809/continental-artificial-intelligence-strategy - Republic of South Africa, Constitution of the Republic of South Africa, 1996, Chapter 2 (Bill of Rights): https://www.justice.gov.za/constitution/SAConstitution-web-eng.pdf --- # AI Policy: South Korea (KR) Canonical page: https://aipolicy.ius.uzh.ch/compare/south-korea Last updated: 2026-03-02T15:36:38.388Z ## General Description [The Framework Act](https://www.law.go.kr/법령/인공지능발전과신뢰기반조성등에관한기본법) on the Development of Artificial Intelligence and the Creation of a Foundation for Trust (AI Framework Act, originally Law No, 20676) is the Republic of Korea's first comprehensive legislative framework governing artificial intelligence (AI). Officially promulgated on January 21, 2025, and effective January 22, 2026, it makes South Korea the second country in the world—after the European Union—and the first in the Asia-Pacific region to adopt a holistic AI law. The Act was subsequently amended on January 20, 2026 (Law No. 21311), with the amended provisions entering into force on July 21, 2026. The amendment principally restructures the national AI governance body by renaming the "National AI Committee" to the "National AI Strategy Committee" (Art. 7), broadening its mandate and institutional capacity, while also legislating protections for AI-vulnerable groups (Art. 3(5)), reinforcing the support framework for AI training data management (Art. 15), venture investment (Art. 18), and AI research institutes (Arts. 22b and 22c). The Act is designed to balance two objectives that can pull in opposite directions: accelerating Korea's national competitiveness in AI while protecting the rights, dignity, and safety of citizens (Art. 1). Rather than following a purely restrictive model, it weaves together promotional policies (R&D investment, startup support, talent cultivation) with explicit obligations for transparency, safety, and accountability—especially for AI systems with the higher potential for societal impact. The AI Framework Act delegates numerous specific matters to [Presidential Decree](https://www.law.go.kr/법령/인공지능발전과신뢰기반조성등에관한기본법시행령/), which serves as the legally binding implementing instrument that gives operational specificity to the Act's framework obligations. Enacted on January 21 and in force since January 22, 2026, the Decree (No. 36053) was promulgated in close coordination with the Act to ensure seamless enforcement from day one. In particular, Chapter 4, titled ‘Securing Artificial Intelligence Ethics and Trustworthiness’, prescribes as follows: - Article 21 – Establishment and Proclamation of AI Ethical Principles - Article 22 – Support for AI Safety and Trustworthiness Verification and Certification - Article 23 – Obligation to Ensure AI Transparency - Article 24 – Obligation to Ensure AI Safety - Article 25 – Procedures for Confirming High-Impact AI - Article 26 – Establishment and Operation of the Expert Committee - Article 27 – Responsibilities of Businesses Related to High-Impact AI - Article 28 – Impact Assessment of High-Impact AI - Article 29 – Criteria for Businesses Required to Designate a Domestic Representative ## Steps Taken Ahead of the January 22, 2026 enforcement of the Enforcement Decree, on November 12, 2025, the Ministry of Science and ICT (MSIT) issued an [Advance Notice for the Enforcement Decree of the AI Framework Act](https://www.msit.go.kr/bbs/view.do?sCode=user&mId=307&mPid=208&bbsSeqNo=94&nttSeqNo=3186786). The Advance Notice for the Enforcement Decree carries significant weight, as it gives concrete and operational form to the obligations that the Act itself articulates only in broad terms, most notably the duties to ensure transparency and safety, the criteria for determining whether an AI system qualifies as high-impact AI, and the corresponding responsibilities of AI business operators. The Ministry of Science and ICT (MSIT) released [five AI Framework Act guidelines](https://www.nia.or.kr/site/nia_kor/ex/bbs/View.do?cbIdx=99835&bcIdx=28987&parentSeq=28987)—including the Guideline on Securing AI Transparency, AI Safety Assurance Guideline, High-Impact AI Determination Guideline, High-Impact AI Business Operator Obligations Guideline, and AI Impact Assessment Guideline. Among these, the transparency guideline addresses the AI Framework Act requirement that businesses providing products or services utilising generative or high-impact AI must notify users in advance of such use—an obligation referred to as the duty to ensure transparency. Issued to resolve concerns that the application criteria were ambiguous, the guidance sets out practical implementation standards and clarifies in particular that this obligation applies only to “AI businesses” that directly provide AI products or services to end users, and not to upstream developers or intermediaries lacking direct user interaction. The MSIT announced a regulatory grace period of at least one year — and potentially longer — before enforcement of the obligations under the AI Framework Act and its Enforcement Decree will be actively applied against businesses. This measure was introduced to minimise disruption to industry and to allow sufficient preparation time for affected operators. During the grace period, a guidance phase will be operated in relation to fact-finding investigations and the imposition of administrative fines. Fact-finding investigations will be conducted only in highly exceptional circumstances, such as where a serious social harm has occurred, for example, loss of life or violation of fundamental human rights. In parallel, MSIT will establish and operate a dedicated [AI Framework Act Support Desk](https://www.sw.or.kr/AI_act_helpdesk/main.jsp) to assist businesses in achieving smooth and practical compliance. The Support Desk will provide specific, practice-oriented advice on matters arising under both the AI Framework Act and its Enforcement Decree. ## Approaches The [AI Framework Act](https://www.law.go.kr/%EB%B2%95%EB%A0%B9/%EC%9D%B8%EA%B3%B5%EC%A7%80%EB%8A%A5%20%EB%B0%9C%EC%A0%84%EA%B3%BC%20%EC%8B%A0%EB%A2%B0%20%EA%B8%B0%EB%B0%98%20%EC%A1%B0%EC%84%B1%20%EB%93%B1%EC%97%90%20%EA%B4%80%ED%95%9C%20%EA%B8%B0%EB%B3%B8%EB%B2%95/\(20676,20250121\)) consists of two primary pillars: (1) a national support system for AI technology and industry promotion, and (2) regulatory obligations for AI businesses, particularly those dealing with “high-impact AI” and “generative AI.” While the act has been welcomed as a crucial legal foundation to facilitate government-backed AI investments, concerns remain about regulatory uncertainty and its potential effects on the business environment. During the legislative process, proposals were made to require AI companies to disclose training datasets and provide transparency for copyright holders, but these were ultimately not included in the final version of the law. It is anticipated that these issues may be addressed through future subordinate legislation or separate laws. And the specific regulatory details of the AI Framework Act will be further refined through enforcement decrees and guidelines, which will shape its practical impact. As a result, AI businesses must closely monitor forthcoming regulations and actively engage in the policy-making process to ensure regulatory clarity and balanced implementation. Specifically, the AI Framework Act establishes a horizontal regulatory framework, where a single agency oversees uniform regulations for “high-impact” AI systems. While this approach reduces regulatory burdens and penalties compared to the EU AI Act, aligning more closely with Canada’s proposed AI and Data Act, it fails to account for the diversity of AI applications. By applying a one-size-fits-all regulatory model, the framework risks overregulation and conflicts with sector-specific regulations. A major challenge of this approach is South Korea’s dual-layered regulatory structure, combining horizontal AI regulations with sector-specific rules. Unlike other countries such as the U.S., the U.K., and Japan, which have opted for sector-specific AI governance, South Korea has integrated AI regulations into multiple ministries' existing regulatory frameworks. This could lead to overlapping jurisdictional conflicts between ministries, particularly as agencies such as the Korea Communications Commission, the Personal Information Protection Commission, and the Financial Services Commission are already developing their own AI-related regulations. To prevent regulatory fragmentation and excessive oversight, the National AI Commission must play a strong central coordinating role to ensure policy consistency. ## Content The Act contains key definitions regarding e.g., "AI"; "High Impact AI"; "Generative AI" etc. Furthermore, it sets out the scope of its application. The Act will apply to any acts conducted abroad that affect the domestic market or users in the Republic of Korea. However, the AI developed or used solely for national defense or national security purposes as prescribed by Presidential Decree is explicitly excluded (Article 4(2) [AI Framework Act](https://www.law.go.kr/%EB%B2%95%EB%A0%B9/%EC%9D%B8%EA%B3%B5%EC%A7%80%EB%8A%A5%20%EB%B0%9C%EC%A0%84%EA%B3%BC%20%EC%8B%A0%EB%A2%B0%20%EA%B8%B0%EB%B0%98%20%EC%A1%B0%EC%84%B1%20%EB%93%B1%EC%97%90%20%EA%B4%80%ED%95%9C%20%EA%B8%B0%EB%B3%B8%EB%B2%95/\(20676,20250121\))). In addition the Act stipulates obligations of AI business operators to ensure transparency and safety of AI. This requires operators e.g. to inform users that generative AI has been used to generate any form of content (such as voices, images, or videos; Article 31 AI Framework Act). Throughout the artificial intelligence life cycle, certain operators must identify, assess, and mitigate risks and must submit results of implementation to the Minister of Science and ICT. ## Sources - AI Framework Act (Korean), scheduled to be enforced on 22 January 2026: https://www.law.go.kr/LSW/lsInfoP.do?lsiSeq=268543&ancYd=20250121&ancNo=20676&efYd=20260122&nwJoYnInfo=N&efGubun=Y&chrClsCd=010202&ancYnChk=0#0000 - AI Safety Institute, 27 November 2024: https://www.aisi.re.kr/eng - Human Rights Impact Assessment (HRIA) Tool, 8 July 2024: https://www.humanrights.go.kr/download/BASIC_ATTACH?storageNo=10005586 - The Seoul Declaration for Safe, Innovative, and Inclusive AI, 21 May 2024: https://www.mofa.go.kr/viewer/skin/doc.html?fn=20240522103241953.pdf&rs=/viewer/result/202503 - Public Official Election Act, 8 March 2024: https://elaw.klri.re.kr/kor_service/lawView.do?hseq=64713&lang=KOR - General Act on Public Administration, 16 January 2024: https://elaw.klri.re.kr/kor_service/lawView.do?hseq=64683&lang=KOR - Personal Information Protection Act, 14 March 2023: https://elaw.klri.re.kr/kor_service/lawView.do?hseq=62389&lang=KOR - the National Guidelines for AI Ethics, 23 December 2020: https://ai.kisdi.re.kr/eng/main/contents.do?menuNo=500010 - National Strategy for Artificial Intelligence, 2019: https://www.msit.go.kr/eng/bbs/view.do?sCode=eng&mId=10&mPid=9&bbsSeqNo=46&nttSeqNo=9 --- # AI Policy: Switzerland (CH) Canonical page: https://aipolicy.ius.uzh.ch/compare/switzerland Last updated: 2025-06-30T07:59:41.373Z ## General Description Switzerland will not adopt a specific regulation on AI, but will adapt the existing legal framework to meet the challenges caused by AI. Based on its signing and expected ratification of the Council of Europe’s AI Convention, Switzerland will publish a legislative proposal by the end of 2026 that will specify the proposed adaptations and amendments to existing laws, focusing on transparency, data protection, non-discrimination, and supervision. The regulatory approach promoted by the federal government is based on three objectives: fostering innovation, protecting fundamental rights (including the right to economic freedom), and enhancing public trust in AI. Key points of the Swiss approach are: - Incorporation of the [Council of Europe's AI Convention](https://www.coe.int/en/web/artificial-intelligence/the-framework-convention-on-artificial-intelligence) into Swiss law. - Focus on amending sector-specific laws, with some adaptations of horizontal laws such as data protection laws. - Non-binding measures (e.g., self-declaration agreements, industry solutions) to support the implementation. ## Steps Taken - The Federal Office of Justice (FOJ), the Federal Office of Communications (OFCOM) and the Federal Department of Foreign Affairs (FDFA) will draft a legislative proposal by the end of 2026 focusing on the amendments needed for the implementation of the Council of Europe's AI Convention, namely regarding transparency, data protection, non-discrimination, and supervision. - On 27 March 2025, Switzerland signed the [AI Convention of the Council of Europe.](https://www.coe.int/en/web/artificial-intelligence/the-framework-convention-on-artificial-intelligence) - On 12 February 2025 the Swiss Federal Council decided on the parameters of a Swiss legal framework on AI. They are: (i) incorporation of the [AI Convention of the Council of Europe](https://www.coe.int/en/web/artificial-intelligence/the-framework-convention-on-artificial-intelligence) into Swiss law; (ii) adaption of the existing legal order with a focus on sector-specific regulations and limited horizontal rules, e.g. in data protection law; (iii) development of non-legally binding measures such as self-disclosure declarations and industry solutions. - In 2024 the Federal Office of Justice (FOJ) wrote a [basic legal analysis](https://www.bakom.admin.ch/dam/bakom/de/dokumente/KI/analyse_juritisch.pdf.download.pdf/Juristische%20Analyse.pdf) on the requirements of the Council of Europe’s AI Convention of the impact of the EU AI Act on Switzerland and the Federal Office of Communications (OFCOM) wrote a [sectoral analysis](https://www.bakom.admin.ch/dam/bakom/de/dokumente/KI/ueberblick_regulierungsvorhaben.pdf.download.pdf/Sektorielle%20Analyse.pdf) providing an overview of the existing and planned adaptations of Swiss federal law in various sectors in connection with AI and a [comparative law report](https://www.bakom.admin.ch/dam/bakom/en/dokumente/KI/analyse_laender.pdf.download.pdf/Analysis%20of%20countries%20and%20regions%20around%20the%20world.pdf) on the regulation of AI in 20 selected jurisdictions. All three reports were published on 12 February 2025. - In 2022 the Swiss Federal Council created the Competence Network for AI ([CNAI](https://cnai.swiss/ueber-das-cnai/)) to facilitate networking on AI projects within the Federal Administration. - On 25 November 2020 the Swiss Federal Council adopted the [Guidelines on Artificial Intelligence for the Confederation.](https://www.sbfi.admin.ch/dam/sbfi/en/dokumente/2021/05/leitlinien-ki.pdf.download.pdf/leitlinien-ki_e.pdf) - On 13 December 2019 the IDAG KI published a Report to the Federal Council on [Challenges of Artificial Intelligence.](https://www.sbfi.admin.ch/dam/sbfi/en/dokumente/2021/05/challenges-ki.pdf.download.pdf/challenges-ki_e.pdf) - In 2019 the Swiss Federal Council established the Interdepartmental Working Group on AI (IDAG KI). ## Approaches According to the federal government, Switzerland will adopt an innovation-friendly, technology-neutral and sector-specific approach to AI regulation. Switzerland will thus not adopt a horizontal regulation of AI maintaining the well-established approach of technology-neutral laws. Instead, Switzerland will selectively adapt the existing legal framework, focusing on sector-specific regulations, such as the Law on Medical Products and Medical Devices and the Road Traffic Act and its ordinance. In addition, some horizontally applicable laws will need to be amended, namely the Data Protection Act and the Product Liability Act. ## Content Switzerland has not yet adopted any adaptation of the Swiss legal framework to meet the challenges caused by AI. A legislative proposal to be published by the end of 2026 will specify the proposed adaptations and amendments to existing laws, focusing on transparency, data protection, non-discrimination, and supervision. ## Sources - Overview of artificial intelligence regulation – Report to the Federal Council, 12 February 2025: https://www.bakom.admin.ch/dam/bakom/en/dokumente/KI/Auslegeordnung%20zur%20Regulierung%20von%20%20k%C3%BCnstlicher%20Intelligenz_def.pdf.download.pdf/Overview.pdf - Overview of Current Sectoral Regulatory Activities Related to Artificial Intelligence – Federal Office of Communications (OFCOM) Report, 16.12.2024: https://www.bakom.admin.ch/dam/bakom/de/dokumente/KI/ueberblick_regulierungsvorhaben.pdf.download.pdf/Sektorielle%20Analyse.pdf - Analysis of artificial intelligence regulation in countries and regions around the world, 16.12.2024: https://www.bakom.admin.ch/dam/bakom/en/dokumente/KI/analyse_laender.pdf.download.pdf/Analysis%20of%20countries%20and%20regions%20around%20the%20world.pdf - Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law, 5 September 2024: https://www.coe.int/en/web/artificial-intelligence/the-framework-convention-on-artificial-intelligence - Legal Basis Analysis in the Context of the Regulatory Framework for Artificial Intelligence, 31.08.2024: https://www.bakom.admin.ch/dam/bakom/de/dokumente/KI/analyse_juritisch.pdf.download.pdf/Juristische%20Analyse.pdf - Guidelines on Artificial Intelligence for the Confederation, 25 November 2020: https://www.sbfi.admin.ch/dam/sbfi/en/dokumente/2021/05/leitlinien-ki.pdf.download.pdf/leitlinien-ki_e.pdf - Challenges of Artificial Intelligence - Executive Summary, 13 December 2019: https://www.sbfi.admin.ch/dam/sbfi/en/dokumente/2021/05/challenges-ki.pdf.download.pdf/challenges-ki_e.pdf --- # AI Policy: Taiwan (TW) Canonical page: https://aipolicy.ius.uzh.ch/compare/taiwan Last updated: 2026-04-27T07:29:08.517Z ## General Description Taiwan is actively developing regulations for artificial intelligence (AI). Initially, the government fostered innovation by easing regulatory constraints, encouraging industries to explore emerging technologies. However, as AI applications have expanded, so too have the associated risks. This shift has prompted the government to place greater emphasis on building an ethical and legal framework for AI. A pivotal moment in this evolution was the 2022 judgment by the Taiwan Constitutional Court in the National Health Insurance Research Database case, which heightened the focus on data governance. Influenced by international developments such as the GDPR and the EU AI Act, some legal scholars advocate for enhanced security and risk management while promoting inclusivity and diversity. This regulatory trajectory entered a new phase when the Artificial Intelligence Basic Act came into force on January 14, 2026. Even so, the Act should not be overstated. It is better understood as a foundational, largely principle-based statute than as a fully operational risk-based regulatory regime. In summary, Taiwan’s approach to AI governance currently combines framework legislation, sectoral guidance, and incremental regulatory initiatives, while still placing primary emphasis on innovation, broad governance principles, and personal data protection rather than on stringent ex ante controls. ## Steps Taken - After the National Science and Technology Council (NSTC) released a draft of the “AI Basic Act” in July 2024 and submitted it to the Executive Yuan in October 2024, the Legislative Yuan passed the bill, and the [Artificial Intelligence Basic Act](https://law.moj.gov.tw/ENG/LawClass/LawAll.aspx?pcode=H0160093) was promulgated and entered into force on January 14, 2026. The Act designates the NSTC as the central competent authority at the national level, requires the Executive Yuan to establish a National AI Strategic Committee, and provides the basic legal framework for subsequent sector-specific AI governance. - The “[Guidelines for Use of Generative AI by the Executive Yuan and Its Subordinate Agencies](https://www.ey.gov.tw/Page/448DE008087A1971/40c1a925-121d-4b6b-8f40-7e9e1a5401f2)” were announced on August 31, 2023, emphasizing responsible and trustworthy use of generative AI in government and highlighting the importance of security, privacy and data governance, and accountability. - The "[AI Technology R&D Guidelines](https://www.nstc.gov.tw/nstc/attachments/53491881-eb0d-443f-9169-1f434f7d33c7)", issued by the Ministry of Science and Technology (MOST) on 23 September 2019, promote adherence to fairness, safety, privacy protection, transparency, explainability, and accountability to foster a human-centric AI society. - The "[AI Taiwan Action Plan (2018-2021)](https://www.twaicoe.org/media/Downloads/%E8%87%BA%E7%81%A3AI%E8%A1%8C%E5%8B%95%E8%A8%88%E7%95%AB.pdf)" implemented by the Executive Yuan, aimed to ease restrictions on innovative technologies and adjust relevant regulations. Following this, the "[AI Taiwan Action Plan 2.0 (2023 - 2026)](https://digi.nstc.gov.tw/File/7C71629D702E2D89/e8ccec35-9e42-431c-b778-45dae073d5b5?A=C)" places greater emphasis on the ethical and legal aspects of AI, aligning with international standards. ## Approaches Taiwan’s approach to AI regulation has been gradual, pragmatic, and innovation-sensitive. Rather than beginning with a comprehensive and highly prescriptive statute, Taiwan initially relied on policy plans, administrative guidelines, and sector-specific experimentation to respond flexibly to rapid technological change. In this respect, Taiwan’s earlier trajectory resembled developments in jurisdictions such as the United States, Japan, and the United Kingdom, where soft-law instruments often preceded comprehensive legislation. The enactment of the Artificial Intelligence Basic Act in 2026, however, marked an important transition. Taiwan’s model is no longer accurately described as “guideline first, legislation later.” It is now better understood as a hybrid governance model that combines framework legislation, administrative guidance, and sector-specific regulation. The Basic Act establishes overarching principles and institutional direction, while leaving many operational and enforcement questions to competent authorities in particular sectors. Substantively, Taiwan’s regulatory approach continues to emphasize a balance between innovation and risk governance. Key concerns include privacy and data governance, cybersecurity, transparency, fairness, accountability, and the protection of fundamental rights. At the same time, Taiwan has so far refrained from adopting a highly restrictive ex ante licensing model or a dense cross-sector compliance regime comparable to the European Union’s approach. More broadly, human well-being, democratic values, and sustainable development remain important normative reference points in Taiwan’s evolving AI governance framework. ## Content Legal reforms are underway in various fields in response to the growing use of AI. Regarding autonomous driving, the [Unmanned Vehicles Technology Innovative Experimentation Act](https://law.moj.gov.tw/ENG/LawClass/LawAll.aspx?pcode=J0030147) was published in December 2018. Under specific conditions approved by the competent authority, operators can enjoy lower regulatory scrutiny and exemptions during the experimental period. In finance, as Taiwan's government intends to encourage fintech services, the [Financial Technology Development and Innovation Experimentation Act](https://law.moj.gov.tw/ENG/LawClass/LawAll.aspx?pcode=G0380254) was published in January 2018. The Act establishes a financial regulatory sandbox designed to foster fintech innovation by providing a controlled environment in which operators may test AI-related services. During the experimentation period, operators can benefit from lower regulatory oversight and exemptions from existing laws lasting up to three years. In addition, the Guidelines for Securities Investment Consulting Services of SICEs with Automated Tools was amended in 2022. It requires financial institutions to adhere to standardized processes and regulatory norms when utilizing AI for robo-advisor investment services. In 2024, the Financial Supervisory Commission (FSC) announced "[Guidelines for the Application of Artificial Intelligence (AI) in the Financial Industry.](https://law.fsc.gov.tw/Download.ashx?FileID=23867&id=2690&type=ANNE)” The guidelines promote flexible risk management, allowing institutions to choose cost-effective mitigation measures based on specific AI scenarios. They emphasize clear risk management responsibilities with third-party providers, encourage independent third-party reviews if needed, and require professional risk control over third-party generative AI outputs. The "data minimization principle" ensures only necessary data is collected. Institutions must explain their AI systems' logic and enhance transparency by disclosing relevant information when appropriate. In health care, amendments to the [Act for the Development of Biotech and Pharmaceutical Industry](https://law.moj.gov.tw/ENG/LawClass/LawAll.aspx?pcode=J0040046) in 2021 expanded its scope to include the digital healthcare industry, granting it tax incentives in line with the biomedical sector. In 2022, amendments to the Medical Act were made to further govern telemedicine applications. Approved AI systems may now be used to assist in remote diagnosis, particularly in under-resourced rural areas. In January 2026, Taiwan formally enacted the [Artificial Intelligence Basic Act](https://law.moj.gov.tw/ENG/LawClass/LawAll.aspx?pcode=H0160093&utm_source=chatgpt.com), transforming what had previously been a policy proposal into a binding framework statute. The enactment of such a framework law had become increasingly important. Without a higher-level legal framework, Taiwan’s AI development could be constrained by the lack of a clearer legal basis and policy direction. At the same time, as AI governance rapidly emerged as an international regulatory issue, Taiwan also faced growing pressure to articulate its own legal framework so as to remain connected to evolving international standards and conversations. In this sense, the Act reflects not only domestic governance concerns but also a broader effort to avoid regulatory marginalization. Substantively, however, the Act remains a framework-oriented and largely principle-based statute rather than a dense regulatory code. It defines AI broadly and sets out seven guiding principles for AI research, development, and application, including human autonomy, privacy protection and data governance, cybersecurity and safety, transparency and explainability, fairness and non-discrimination, and accountability. It also seeks to promote innovation, digital equity, social well-being, cultural values, and international competitiveness. The Act refers to AI risk classification, high-risk AI, and follow-up sector-specific regulation, but these mechanisms should not be overstated. Although the Ministry of Digital Affairs is instructed to promote an AI risk taxonomy and assessment framework, and to assist sectoral regulators in establishing risk-based management rules, the statute itself provides only a relatively general structure and leaves most of the operational details for later development. Likewise, while the Act requires the government to review and revise inconsistent laws, regulations, and administrative measures within two years of its entry into force, the practical impact of this clause remains uncertain. It is therefore more accurate to understand the AI Basic Act as establishing the legal and policy vocabulary of AI governance in Taiwan, rather than as already creating a robust and fully operational risk-based regulatory regime. In sum, Taiwan has moved from a predominantly guideline-based model of AI governance to a hybrid model that combines framework legislation, sector-specific regulation, and administrative guidance. ## Sources - Artificial Intelligence Basic Act, 14 January 2026: https://law.moj.gov.tw/ENG/LawClass/LawAll.aspx?pcode=H0160093&utm_source=chatgpt.com - Guidelines for the Application of Artificial Intelligence (AI) in the Financial Industry, 20 June 2024: https://law.fsc.gov.tw/Download.ashx?FileID=23867&id=2690&type=ANNE - Guidelines for Use of Generative AI by the Executive Yuan and Its Subordinate Agencies, 31 August 2023: https://www.ey.gov.tw/Page/448DE008087A1971/40c1a925-121d-4b6b-8f40-7e9e1a5401f2 - AI Taiwan Action Plan 2.0, February 2023: https://digi.nstc.gov.tw/File/7C71629D702E2D89/e8ccec35-9e42-431c-b778-45dae073d5b5?A=C - Act for the Development of Biotech and Pharmaceutical Industry, 30 December 2021: https://law.moj.gov.tw/ENG/LawClass/LawAll.aspx?pcode=J0040046 - AI Technology R&D Guidelines, 23 September 2019: https://www.nstc.gov.tw/folksonomy/detail/dbf8da09-22be-4ef1-8294-8832fc6e8a26?l=ch - Unmanned Vehicles Technology Innovative Experimentation Act, 19 December 2018: https://law.moj.gov.tw/ENG/LawClass/LawAll.aspx?pcode=J0030147 - Financial Technology Development and Innovation Experimentation Act, 31 January 2018: https://law.moj.gov.tw/ENG/LawClass/LawAll.aspx?pcode=G0380254 --- # AI Policy: United Kingdom (GB) Canonical page: https://aipolicy.ius.uzh.ch/compare/united-kingdom Last updated: 2025-06-30T07:56:44.362Z ## General Description The UK does not have a comprehensive AI regulation. The focus lies on fostering economic growth and public sector transformation, with targeted regulation reserved for the most powerful AI systems. Key milestones, such as the 2023 AI Safety Summit and the 2025 [AI Opportunities Action Plan](https://www.gov.uk/government/publications/ai-opportunities-action-plan-government-response), underscore a commitment to balancing opportunity with safety, guided by five core principles: safety, transparency, fairness, accountability, and redress. ## Steps Taken - In March 2025, [Artificial Intelligence (Regulation) Bill](https://bills.parliament.uk/bills/3942) was reintroduced and has passed the first reading in the House of the Lords. - In February 2025 the Department for education published [guidance](https://www.gov.uk/government/publications/generative-ai-product-safety-expectations) that targets developers and suppliers of educational technology on safety expectations of such tools and the Cabinet Office released [guidance](https://www.gov.uk/government/publications/ai-playbook-for-the-uk-government) for civil servants on the use of AI. - In February 2025 a Private Members’ Bill (starting in the House of Lords), the [Public Authority Algorithmic and Automated Decision-Making Systems Bill](https://bills.parliament.uk/bills/3760) was introduced. It is currently at its third reading in the House of Lords. The bill seeks to regulate the use of automated and algorithmic tools in decision-making to the public sector; to require public authorities to complete an impact assessment of automated and algorithmic decision -making systems; to ensure the adoption of transparency standards for such systems; and for connected purposes. - In January 2025 the Secretary of State for Science, Innovation and Technology presented the [AI Opportunities Action Plan](https://www.gov.uk/government/publications/ai-opportunities-action-plan) that argues for an pro-innovation approach and advocates for regulation only where necessary to shield citizens from the gravest risks associated with the most advanced AI systems . In its [response to the Consultation Outcome](https://www.gov.uk/government/consultations/ai-regulation-a-pro-innovation-approach-policy-proposals/outcome/a-pro-innovation-approach-to-ai-regulation-government-response), the government overall agrees with the recommended approach, hinting at binding measures for “highly capable” AI systems. - In July 2024 his Majesty King Charles III [announced](https://www.gov.uk/government/speeches/the-kings-speech-2024) plans for targeted legislation on AI. - Following the AI Safety Summit, in late November 2023, Lord Holmes of Richmond introduced the Artificial Intelligence (Regulation) Bill. However, it did not proceed due the announcement on 22 May 2024 of a general election and the subsequent prorogation of Parliament. - On 1 and 2 November 2023 the AI Safety Summit was hosted in London. The global summit secured voluntary safety commitments from AI firms and 28 states signed the [Bletchley Declaration.](https://www.gov.uk/government/publications/ai-safety-summit-2023-the-bletchley-declaration/the-bletchley-declaration-by-countries-attending-the-ai-safety-summit-1-2-november-2023) - A 2023 [white paper](https://www.gov.uk/government/publications/ai-regulation-a-pro-innovation-approach) set out a principles-based AI framework, relying on existing regulators to oversee AI without new laws. - [](https://www.gov.uk/government/publications/national-ai-strategy)In 2021 the Secretary of State for Digital, Culture, Media and Sport presented a [National AI Strategy](https://www.gov.uk/government/publications/national-ai-strategy) which outlined innovation goals and flagged need for governance, setting a foundation for future regulation. ## Approaches [The King’s Speech of July 2024](https://www.gov.uk/government/speeches/the-kings-speech-2024) under the new Labour government confirmed plans for targeted legislation focused on the most powerful AI models. This approach was affirmed and refined in the [AI Opportunites Action Plan](https://www.gov.uk/government/publications/ai-opportunities-action-plan) that seeks to put the United Kingdom's focus on innovation and economic growth, with regulation limited to the "gravest risks" for citizens and assets of the most advanced AI systems. ## Content The framework laid out in the [white paper](https://www.gov.uk/government/publications/ai-regulation-a-pro-innovation-approach/white-paper#ministerial-foreword) is underpinned by five principles for the responsible use and development of AI in all sectors of economy. - **Safety, security, and robustness:** Throughout the AI life cycle all AI systems should function in a robust, secure, and safe way. Risk must be assessed continually. Safety is a core consideration for the health sector and critical infrastructure and will also play an important role in all other sectors. Furthermore, AI systems must be technically secure and should reliably function as they were intended to. - **Appropriate transparency and explainability:** Information for how, when, and for which purposes an AI system is being used should be transparent. Explainability means the extent to which it is possible to access, interpret and understand the decision-making process of an AI system. The degree of transparency and explainability will both be dependent on the risks posed by the respective system. This is to ensure public trust, which is one of the significant drivers of AI adoption. - **Fairness:** AI systems must not discriminate unfairly or create unfair market outcomes. Legal rights of individuals and organisations must be respected. Fairness as reflected in many areas of law (equality and human rights, data protection, consumer and competition law, public and common law, and rules protecting vulnerable people), should always be adhered to. Certain decisions that will be made by AI based on e.g., credit scores, health data and application documents will greatly influence individuals. Therefore, decisions made by AI systems must always be justifiable and in accordance with the law applicable. - **Accountability and governance:** Since AI systems can operate with a high degree of autonomy and not all decisions made can be foreseen, it must be clear who is accountable for the decisions made. Moreover, proper governance should be established. For example, by providing documentation on key decisions or by conducting impact assessments. - **Contestability and redress:** If rights are violated, people and organisations should be able to contest outcomes. The [Artificial Intelligence (Regulation) Bill](https://bills.parliament.uk/bills/3942) seeks to create an "AI Authority Body" that oversees regulators and the general adherence to the five principles laid out in the white paper. It also foresees the construction of regulatory sandboxes for AI. ## Sources - Artificial Intelligence (Regulation) Bill, reintroduced 4 March 2025: https://bills.parliament.uk/bills/3942 - UK Government's consultation on copyright and artificial intelligence – ICO Response, 25 February 2025: https://ico.org.uk/about-the-ico/consultations/uk-governments-consultation-on-copyright-and-artificial-intelligence/ - • Public Authority Algorithmic and Automated Decision-Making Systems Bill [HL] - Parliamentary Bills – UK Parliament, 10 February 2025: https://bills.parliament.uk/bills/3760 - AI Playbook for the UK Government – Guidance, 10 February 2025: https://www.gov.uk/government/publications/ai-playbook-for-the-uk-government - Generative AI: product safety expectations – Guidance, 22 January 2025: https://www.gov.uk/government/publications/generative-ai-product-safety-expectations - AI Opportunities Action Plan: government response – Policy paper, 13 January 2025: https://www.gov.uk/government/publications/ai-opportunities-action-plan-government-response - AI Opportunities Action Plan – Independent report, 13 January 2025: https://www.gov.uk/government/publications/ai-opportunities-action-plan - Copyright and Artificial Intelligence, 17 December 2024: https://www.gov.uk/government/consultations/copyright-and-artificial-intelligence/copyright-and-artificial-intelligence#c-our-proposed-approach - The King's Speech 2024, 17 July 2024: https://www.gov.uk/government/speeches/the-kings-speech-2024 - New principles on use of AI in education, 4 July 2024: https://russellgroup.ac.uk/news/new-principles-on-use-of-ai-in-education/ - Regulating AI: the ICO's strategic approach - a response to the DSIT Secretary of State, 30 April 2024: https://ico.org.uk/about-the-ico/consultations/regulating-ai-the-icos-strategic-approach-a-response-to-the-dsit-secretary-of-state/ - A pro-innovation approach to AI regulation: government response – Consultation Outcome, 6 February 2024: https://www.gov.uk/government/consultations/ai-regulation-a-pro-innovation-approach-policy-proposals/outcome/a-pro-innovation-approach-to-ai-regulation-government-response#summary-of-consultation-evidence-and-government-response - The governance of artificial intelligence: interim report (9th version), 31 August 2023: https://publications.parliament.uk/pa/cm5803/cmselect/cmsctech/1769/report.html - A pro-innovation approach to AI regulation – Policy Paper, 3 August 2023: https://www.gov.uk/government/publications/ai-regulation-a-pro-innovation-approach/white-paper#ministerial-foreword - Generative artificial intelligence (AI) in education – Policy paper, 29 March 2023: https://www.gov.uk/government/publications/generative-artificial-intelligence-in-education - National AI Strategy, 22 September 2021: https://www.gov.uk/government/publications/national-ai-strategy --- # AI Policy: United States of America (US) Canonical page: https://aipolicy.ius.uzh.ch/compare/united-states-of-america Last updated: 2025-06-30T13:32:42.136Z ## General Description The United States has not enacted comprehensive federal AI legislation. While Congress has introduced multiple bills, none have been enacted into law, leaving AI regulation to a patchwork of state- and city-level laws, as well as department-, agency-, and sector-specific policies and guidelines. With its [Executive Order 14110](https://www.federalregister.gov/documents/2023/11/01/2023-24283/safe-secure-and-trustworthy-development-and-use-of-artificial-intelligence) former President Biden’s administration had set a trajectory for an AI policy that focused on safe and trustworthy AI. After President Trump’s revocation of the Executive Order and the issuance of a much shorter replacement, it is uncertain how AI policy will evolve. ## Steps Taken - On 24 April 2025, President Trump signed [Executive Order 14277](https://www.federalregister.gov/documents/2025/04/28/2025-07368/advancing-artificial-intelligence-education-for-american-youth) that focuses on AI education for the American youth. - President Trump signed [Executive Order 14179 – Removing Barriers to American Leadership in Artificial Intelligence](https://www.federalregister.gov/documents/2025/01/31/2025-02172/removing-barriers-to-american-leadership-in-artificial-intelligence) on 23 January 2025. Though this order marks a pivotal shift, aiming to enhance American dominance in AI by revoking the Biden administration's [Executive Order 14110 – Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence](https://www.federalregister.gov/documents/2023/11/01/2023-24283/safe-secure-and-trustworthy-development-and-use-of-artificial-intelligence), its implementation is not, as yet, clear. - Even though [Executive Order 14110](https://www.federalregister.gov/documents/2023/11/01/2023-24283/safe-secure-and-trustworthy-development-and-use-of-artificial-intelligence) was rescinded and with it the previous focus on safety, the [U.S. Artificial Intelligence Safety Institute](https://www.nist.gov/aisi) remains active, as seen in [recent NIST updates.](https://csrc.nist.gov/pubs/ai/100/2/e2025/final) - [Executive Order 14110](https://www.federalregister.gov/documents/2023/11/01/2023-24283/safe-secure-and-trustworthy-development-and-use-of-artificial-intelligence), dated 30 October, 2023 focused on safe, secure, and trustworthy AI development. The focus on safety was also evident in the [Blueprint for the AI Bill of Rights](https://bidenwhitehouse.archives.gov/ostp/ai-bill-of-rights/) and in the [Memorandum on Advancing the United States’ Leadership in Artificial Intelligence](https://bidenwhitehouse.archives.gov/briefing-room/presidential-actions/2024/10/24/memorandum-on-advancing-the-united-states-leadership-in-artificial-intelligence-harnessing-artificial-intelligence-to-fulfill-national-security-objectives-and-fostering-the-safety-security/). However, these documents are no longer available on the official White House website. - On 22 January 2025, [Project Stargate](https://group.softbank/en/news/press/20250122), led by Japan-based Softbank and U.S.-based OpenAI and Oracle, announced a $ 500 billion private investment in U.S.-based artificial intelligence infrastructure. - On 14 January 2025 former President Biden signed [Executive Order 14141](https://www.federalregister.gov/documents/2025/01/17/2025-01395/advancing-united-states-leadership-in-artificial-intelligence-infrastructure) to advance the US' leadership in AI infrastructure. This order has not been revoked. Prior to this, the first Trump administration (2017-2021) laid foundational policies through two key executive orders regarding AI: - [Executive Order 13859 – Maintaining American Leadership in Artificial Intelligence](https://www.federalregister.gov/documents/2019/02/14/2019-02544/maintaining-american-leadership-in-artificial-intelligence), signed on 11 February 2019, focusing on maintaining U.S. leadership through investments in AI research, development of AI applications across sectors, and ensuring alignment with American values, emphasizing innovation, industry, workers, and national security. - [Executive Order 13960 – Promoting the Use of Trustworthy Artificial Intelligence in the Federal Government](https://www.federalregister.gov/documents/2020/12/08/2020-27065/promoting-the-use-of-trustworthy-artificial-intelligence-in-the-federal-government). dated 3 December 2020, directed federal agencies to ensure their AI use was trustworthy, focusing on safety, security, privacy, and civil rights. These orders remain in effect, as the 2025 order does not explicitly revoke them, leaving the possibility of continuity with earlier Trump administration priorities. Beyond executive orders, federal departments have developed AI-related policies. For instance, the Department of State has its "[Enterprise Artificial Intelligence – Strategy FY 2024-2025: Empowering Diplomacy through Responsible AI](https://2021-2025.state.gov/wp-content/uploads/2023/11/Department-of-State-Enterprise-Artificial-Intelligence-Strategy.pdf)", focusing on ethical AI design, development, acquisition, and application, with goals like leveraging secure AI infrastructure and ensuring responsible application. On state-level states have enacted various laws. notable examples include: - State of California: - [Senate Bill No. 942 – California AI Transparency Act,](https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202320240SB942) taking effect on 1 January 2026, addressing large providers of generative AI to label AI generated content, offer tools to detect such content and to include hidden "latent disclosure" in AI generated content. - [Assembly Bill No. 2013 – Generative artificial intelligence: training data transparency](https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202320240AB2013), taking effect on 1 January 2026, requiring developers of generative AI systems or services, released or substantially modified on or after January 1, 2022, to publicly disclose on their websites a high-level summary of the datasets used to train their AI. - State of Colorado: [Senate Bill 24-205 – Colorado AI Act](https://leg.colorado.gov/bills/sb24-205), taking effect 1 February 2026, targets "high-risk" AI systems – those making or significantly influencing "consequential decisions" in areas like employment, housing, healthcare, and finance. The act requires developers and deployers to use reasonable care to prevent "algorithmic discrimination" based on protected traits (e.g., race, sex, age). - State of Utah: [Senate Bill 149 – Artificial Intelligence Policy Act](https://le.utah.gov/%7E2024/bills/static/SB0149.html), 1 May 2024, regulating private-sector generative AI use. It requires businesses to disclose when consumers interact with generative AI (e.g., chatbots) if asked, and mandates proactive disclosure for regulated professions like healthcare or accounting. - State of Tennessee: [Ensuring Likeness Voice and Image Security (“ELVIS”) Act (Tenn. Code Ann. §47-25-1101 et seq.)](https://www.capitol.tn.gov/Bills/113/Bill/HB2091.pdf), 1 July 2024, addressing AI-generated deepfakes and voice clones prohibiting their unauthorized commercial use and distribution, including by those who provide tools primarily designed for such purposes. - The states of [Virgina](https://lis.virginia.gov/bill-details/20251/HB2094/text/HB2094), [Illinois](https://www.ilga.gov/legislation/BillStatus.asp?DocNum=3563&GAID=17&DocTypeID=HB&SessionID=112&GA=103), [Texas](https://capitol.texas.gov/BillLookup/History.aspx?LegSess=88R&Bill=HB2060) and many other states have also enacted AI relevant policies. Recent developments on state-level can be tracked on the [NCSL Website.](https://www.ncsl.org/technology-and-communication/artificial-intelligence-2025-legislation) On city level, [New York City's Local Law 144](https://legistar.council.nyc.gov/LegislationDetail.aspx?ID=4344524&GUID=B051915D-A9AC-451E-81F8-6596032FA3F9&Options=ID%7cText%7c&Search=) mandates bias audits and candidate notifications for employment decision tools. In May 2023, the [U.S. National Science Foundation](https://www.nsf.gov/news/nsf-announces-7-new-national-artificial) announced a $ 140 million investment to establish seven new National Artificial Intelligence Research Institutes that contribute to the development of ethical and trustworthy AI systems and technologies. However NSF has pulled funding from many existing projects and new funding is currently frozen. ## Approaches The US approach to AI regulation is decentralized and incremental. So far the US has no unified federal framework. Executive orders aim to promote innovation, while executive departments and agencies tailor guidelines to their needs. States and cities enact targeted laws – addressing discrimination, transparency, and deepfakes – forming a patchwork of specific, reactive policies rather than a comprehensive strategy. ## Content In the US, there has been no comprehensive regulation regarding AI. At the federal level, President Trump has signed [](https://www.federalregister.gov/documents/2025/01/31/2025-02172/removing-barriers-to-american-leadership-in-artificial-intelligence)[Executive Order 14179](https://www.federalregister.gov/documents/2025/01/31/2025-02172/removing-barriers-to-american-leadership-in-artificial-intelligence) – it prioritizes American dominance in AI and at the same time revokes former President Biden’s safety-focused [](https://www.federalregister.gov/documents/2023/11/01/2023-24283/safe-secure-and-trustworthy-development-and-use-of-artificial-intelligence)[Executive Order 14110](https://www.federalregister.gov/documents/2023/11/01/2023-24283/safe-secure-and-trustworthy-development-and-use-of-artificial-intelligence). Additionally, federal department-, agency- and sector-specific guidelines address a range of detailed challenges. This includes, e.g., [department-specific visions](https://2021-2025.state.gov/wp-content/uploads/2023/11/Department-of-State-Enterprise-Artificial-Intelligence-Strategy.pdf) for the use of AI, as well as recommendations on how AI should be implemented in sectors such as [education](https://www.ed.gov/sites/ed/files/documents/ai-report/ai-report.pdf). On the state level, policies in – or about to be – in force represent a decentralized and targeted regulatory patchwork. Although some regulations address multiple issues, most policies focus on specific concerns like training data and origin transparency, algorithmic discrimination, AI interaction disclosure, and deepfake protection. On City level, some AI relevant local laws with a rather narrow scope have been enacted including the use of AI for surveillance. ## Sources - State of Virginia: House Bill 2094 – High-Risk Artificial Intelligence Developer and Deployer Act, if signed, effective 1 July 2026: https://lis.virginia.gov/bill-details/20251/HB2094/text/HB2094 - State of California: Assembly Bill No. 2013 – Generative artificial intelligence: training data transparency, effective 1 January 2026: https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202320240AB2013 - State of California: Senate Bill 942 – California AI Transparency Act, effective 1 January 2026: https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202320240SB942 - State of Colorado: Senate Bill 24-205 – Colorado AI Act, effective 1 February 2026: https://leg.colorado.gov/bills/sb24-205 - Executive Order 14277 DJT – Advancing Artificial Intelligence Education for American Youth, 24 April 2025: https://www.federalregister.gov/documents/2025/04/28/2025-07368/advancing-artificial-intelligence-education-for-american-youth - NIST AI 100-2 E2025 –Adversarial Machine Learning: A Taxonomy and Terminology of Attacks and Mitigations, March 2025: https://csrc.nist.gov/pubs/ai/100/2/e2025/final - Executive Order 14179 DJT – Removing Barriers to American Leadership in Artificial Intelligence, 23 January 2025: https://www.federalregister.gov/documents/2025/01/31/2025-02172/removing-barriers-to-american-leadership-in-artificial-intelligence - Project Stargate, 22 January 2025: https://group.softbank/en/news/press/20250122 - Executive Order 14141 JB – Advancing United States Leadership in Artificial Intelligence Infrastructure, 14 January 2025: https://www.federalregister.gov/documents/2025/01/17/2025-01395/advancing-united-states-leadership-in-artificial-intelligence-infrastructure - Memorandum on Advancing the United States’ Leadership in Artificial Intelligence, 24. October 2024: https://bidenwhitehouse.archives.gov/briefing-room/presidential-actions/2024/10/24/memorandum-on-advancing-the-united-states-leadership-in-artificial-intelligence-harnessing-artificial-intelligence-to-fulfill-national-security-objectives-and-fostering-the - State of Illinois: House Bill 3563 – Illinois AI Task Force Act, 4 August 2024: https://www.ilga.gov/legislation/BillStatus.asp?DocNum=3563&GAID=17&DocTypeID=HB&SessionID=112&GA=103 - State of Tennessee: Ensuring Likeness Voice and Image Security (“ELVIS”) Act (Tenn. Code Ann. §47-25-1101 et seq.), 1 July 2024: https://www.capitol.tn.gov/Bills/113/Bill/HB2091.pdf - State of Utah: Senate Bill 149 – Artificial Intelligence Policy Act, 1 May 2024: https://le.utah.gov/%7E2024/bills/static/SB0149.html - Enterprise Artificial Intelligence – Strategy FY 2024-2025: Empowering Diplomacy through Responsible AI, 9 November 2023: https://2021-2025.state.gov/wp-content/uploads/2023/11/Department-of-State-Enterprise-Artificial-Intelligence-Strategy.pdf - FACT SHEET: Vice President Harris Announces New U.S. Initiatives to Advance the Safe and Responsible Use of Artificial Intelligence, 1 November 2023: https://bidenwhitehouse.archives.gov/briefing-room/statements-releases/2023/11/01/fact-sheet-vice-president-harris-announces-new-u-s-initiatives-to-advance-the-safe-and-responsible-use-of-artificial-intelligence/ - State of Texas: House Bill 2060, September 2023: https://capitol.texas.gov/BillLookup/Text.aspx?LegSess=88R&Bill=HB2060 - U.S. National Science Foundation – Artificial Intelligence Research Institutes, 4 May 2023: https://www.nsf.gov/news/nsf-announces-7-new-national-artificial - New York City's Local Law 144, 1 January 2023: https://legistar.council.nyc.gov/LegislationDetail.aspx?ID=4344524&GUID=B051915D-A9AC-451E-81F8-6596032FA3F9&Options=ID%7cText%7c&Search= - Executive Order 14110 JB – Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence, 30 October 2023: https://www.federalregister.gov/documents/2023/11/01/2023-24283/safe-secure-and-trustworthy-development-and-use-of-artificial-intelligence - Executive Order 13960 DJT – Promoting the Use of Trustworthy Artificial Intelligence in the Federal Government, 3 December 2020: https://www.federalregister.gov/documents/2020/12/08/2020-27065/promoting-the-use-of-trustworthy-artificial-intelligence-in-the-federal-government - Executive Order 13859 DJT – Maintaining American Leadership in Artificial Intelligence, 11 February 2019: https://www.federalregister.gov/documents/2019/02/14/2019-02544/maintaining-american-leadership-in-artificial-intelligence