Timeline of AI policy

From Timelines
Jump to: navigation, search

This is a timeline of AI policy and legislation, which attempts to overview the changes in international and local laws around AI and AI safety.

Full timeline

Year Month and date Region Name Event type Details
2017 June Canada Pan-Canadian Artificial Intelligence Strategy National Policy Canada releases the world’s first national AI strategy, aiming to have the most robust AI ecosystem in the world by 2030.[1] The strategy is a collaborative effort, spanning across government, academia, and industry sectors and headed by the Canadian Institute for Advanced Research (CIFAR).[2] Canada’s names the Vector Institute (Canada), Mila (research institute), and Amii (research institute) as national AI institutes and contributors to the nation’s AI progress.[1] This strategy would go on to enhance Canada’s global standing in AI research and innovation.
2017 July 20 China China Guidelines on AI Development National Policy The State Council of the People's Republic of China issues guidelines on developing AI by embedding AI into the socioeconomic landscape and the country’s basic functioning. The council lays out plans to be a world leader in AI by 2030, aiming for the total output of the AI industry to be 1 trillion yuan ($147.8 billion).[3]
2018 May 5 European Union General Data Protection Regulation International Policy The European Union effects the General Data Protection Regulation (GDPR), the strongest and most comprehensive attempt yet to regulate personal data. The GDPR outlines a set of rules that aims to strengthen protection for personal data in response to increasing development in the tech world.[4] Although the GDPR is focused on privacy, it states that individuals have the right to a human review of results from automated decision-making systems.[5] The fine for violating the GDPR is high and extends to any organization that offers services to EU citizens.[4]
2018 June 28 United States California Consumer Privacy Act Regional Policy The California Consumer Privacy Act is signed into law, heightening consumer control over personal information. The law would go into effect January 1, 2020 and grants consumers the right to know about, opt out of the sharing of, and delete personal information[6]. The Act would influence personal data usage by giving consumers the right to opt out of automated decision-making systems and by compelling businesses to inform customers on how and for what purpose they use personal information[7]. These regulations would require businesses to disclose if and how they use personal information for AI training.
2019 February 11 United States Executive Order 13859 National Policy President Trump signs Executive Order 13859 to maintain American Leadership in Artificial Intelligence. The Order directs federal agencies to prioritize AI research and develop and prompt American leadership in the AI space.[8] The Order does not provide details on how it plans to put the new policies in effect, and does not allocate any federal funding towards executing its vision.[9]
2019 March 29 Japan Social Principles of Human-Centered AI National Policy The Japanese government releases the Social Principles of Human-Centered AI, a set of guidelines for implementing AI in society with the philosophies of human dignity, diversity and inclusion, and sustainability which the government will continuously revise.[10] Japan provides nonbinding guidelines on AI and imposes transparency obligations on some large digital platforms.[11] Japan aims to achieve social goals through the use of AI rather than restriction. The structure of the principles corresponds to the OECD AI Principles, which also outline AI’s potential alongside its risks.[12].
2019 April 10 United States Algorithmic Accountability Act National Policy The Algorithmic Accountability Act bill is introduced into the house. Commercial entities must “conduct assessments of high-risk systems that involve personal information or make automated decisions, such as systems that use artificial intelligence or machine learning.”[13] The Bill aims to minimize bias, discrimination, and inaccuracy in automated decision systems by compelling companies to assess their impacts. The Act does not establish binding regulations but asks the Federal Trade Commission to establish rules for evaluating highly sensitive automated systems.[14] The legislation would be introduced into the senate in 2022[15] but would still not be law through 2024.
2019 May 29 International OECD AI Principles International Policy The Organisation for Economic Co-operation and Development (OECD) issues AI principles to shape policies, create an AI risk framework, and to foster global communication and understanding across jurisdictions. The European Union, Council of Europe, United Nations, and the United States would use these principles in their AI legislation.[16] The principles aim to be values-based and include the following categories: sustainable development, human rights, transparency and explainability, security, and accountability.[17] The principles would be updated again in May 2024 in consideration of new technology and policy developments.[16]
2019 November 7 Australia Australia’s Artificial Intelligence Ethics Framework National Policy The Australian Government releases an artificial intelligence ethics framework to ensure safe, secure, and reliable AI. The framework includes eight voluntary, nonbinding principles to complement existing AI practices: human well-being, human-centered values, fairness, privacy protection, reliability, transparency and explainability, contestability, and accountability.[18] The principles are on par with those set forth by OECD and the World Economic Forum, and set to be trialed by National Australia Bank, Commonwealth Bank, Telstra, Microsoft, and Flamigo AI.[19]
2020 June International Global Partnership on Artificial Intelligence International Organization The Global Partnership on Artificial Intelligence (GPAI) is established to share multidisciplinary research, identify key issues in AI, and facilitate international collaboration.[20] The OECD hosts the GPAI secretariat and the partnership is based on the shared commitment of G7 countries to the OECD AI Principles.[20] The Alliance is a multi-stakeholder initiative to foster international cooperation over AI and includes the working groups Responsible AI, Data Governance, Future of Work, and Innovation and Commercialization.[20] The United States, wary to join any international AI panel due to overregulation concerns, joined the partnership to counter China’s increasing international AI presence.[21]
2020 October 10 Europe Artificial Intelligence Act International Policy European Union leaders meet to discuss the digital transition. They invite the European Commission, the executive branch of the EU, to increase private and public tech investment, ensure elevated coordination between European research centers, and construct a clear definition of Artificial Intelligence[22].
2021 April 21 Europe Artificial Intelligence Act International Policy The European Commission proposes the Artificial Intelligence Act. The Commission releases a proposal for AI regulation aiming to improve trust in AI and foster its development[22].
2021 September 22 United Kingdom UK National AI Strategy National Policy The UK government releases its National AI Strategy - a 10-year plan that outlines how to invest in and plan for long-term AI ecosystem needs, support the transition to an AI-enabled economy, and ensure the UK succeeds in AI governance.[23] This strategy comes a few months after the EU’s Artificial Intelligence Act. The Alan Turing Institute, established in 2015, is the national research center for AI and one of the organizations that will help implement the AI strategy.[24]
2021 September 30 Brazil Brazilian Strategy for Artificial Intelligence National Policy The Brazilian Government approves the Brazilian Strategy for Artificial Intelligence, a document to guide research, innovation, and development of ethical AI solutions.[25] The strategy is based on the OECD AI Principles, aiming to develop ethical AI principles, guide AI use, remove barriers to innovation, improve cross-sector collaboration, develop AI skills, promote AI investment, and advance Brazilian technology overseas.[26] The strategy faced criticism for its lack of specifics on regulation.[27]
2023 June 15 International AI Governance Alliance (AIGA) Established International Organization The World Economic Forum launches the AI Governance Alliance (AIGA) to guide responsible development and deployment of AI systems.[28] The Alliance prioritizes safe systems and technology, promoting sustainable applications and transformation, and contributing to resilient governance and regulation.[29] Its members would be from industry, government, and civil society worldwide.
2023 July 12 United States Whitehouse Meets with AI Companies National Policy President Joe Biden and Vice President Kamala Harris host leading AI companies Amazon (company), Anthropic, Google, Inflection AI, Meta Platforms, Microsoft, and OpenAI at the Whitehouse and secure their voluntary commitments to prioritize safe, secure, and transparent AI development.[30] The companies promise to ensure product safety before public introduction, build secure systems, and earn public trust. Congress has yet to pass contemporary AI bills, so this voluntary, nonbinding agreement is the primary guidance around AI concerns.[31]
2023 September 19 United States Anthropic’s Responsible Scaling Policy Company Policy Anthropic publishes its Responsible Scaling Policy (RSP) - a series of technical and organizational protocols to guide risk management and development of increasingly powerful AI systems.[32] The RSP delineates AI Safety Levels 1-4, loosely based on the US Governments biosafety levels, to address catastrophic risks.[32] Anthropic aims in part to create competition in the AI safety space by publishing the policy.[32] The Institute for AI Policy and Strategy offers critiques of Anthropic’s RSP. The institute states the risk thresholds should be based on absolute risk rather than relative risk, the risk level thresholds should be lower than Anthropic defines them, and Anthropic should outline when it will alert authorities of identified risks and commit to outside scrutiny and evaluations. [33]
2023 October 30 United States Executive Order 14110 National Policy Biden signs Executive Order 14110 titled Executive Order on Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence. The Order establishes new standards for AI safety and security. It compels developers to share test results with the US government and create tools to ensure AI system safety, protects Americans from AI fraud and deception, sets up a cybersecurity program to develop AI tools and fix vulnerabilities, and orders the development of a national security memorandum that directs future AI security measures[34]. The Order also directs the National Institute of Standards and Technology (NIST) to develop standards for evaluation and red-teaming and to provide testing environments for AI systems. The general reaction to the bill is cautious optimism[35]. As Less Wrong blogger Zvi Mowshowitz reports, some worry that this is the first step in a slippery slope of heightened regulation that could dampen innovation and development[36]. A complete timeline and outlook of the Executive Order can be found here.[37]
2023 November 1 – 2 International Conference AI Safety Summit International Policy The first AI Safety Summit is held at Bletchley Park, Milton Keynes in the United Kingdom. It leads to an agreement known as the Bletchley Declaration by the 28 countries participating in the summit, including the United States, United Kingdom, China, and the European Union.[38] It receives some commentary on LessWrong, viewing it as a partial step in the right direction,[39] including a lengthy blog post by Zvi Mowshowitz, a frequent commentator on AI developments from an AI safety lens.[40]
2023 November 1 United States US AI Safety Institute Organization United States Vice President Kamala Harris announces the U.S. AI Safety Institute (USAISI) at the AI Safety Summit in the United Kingdom. The launch of USAISI builds on Biden's executive order of two days ago (October 30).[41]
2023 November 2 United Kingdom UK AI Safety Institute Organization The United Kingdom government announces the launch of the UK AI Safety Institute. The UK AI Safety Institute is to be formed from the Frontier AI Taskforce, which in turn had previously been called the Foundation Model Taskforce. Ian Hogarth serves as its chair.[42]
2023 December 9 Europe Artificial Intelligence Act International Policy The European Council and European Parliment reach a provisional agreement on the Artificial Intelligence Act. The Act should go into effect in 2026.[22]
2023 December 18 United States OpenAI Publishes Preparedness Framework Company Policy OpenAI releases its “Preparedness Framework,” a living document positing that a “robust approach to AI catastrophic risk safety requires proactive, science-based determinations of when and how it is safe to proceed with development and deployment.”[43] The elements of the framework include tracking catastrophic risk with evaluations, seeking out unknown-unknowns, establishing safety baselines, tasking preparedness teams with on-the-ground-work, and creating a cross-functional advisory board.[44] This document is released a few months after Anthropic’s RSP. Safer AI comments on OpenAI’s improvements to Anthropic’s safety document, including calling for more safety tests, allowing the board to veto CEO decisions, adding risk identification and analysis, and forecasting risks.[44] Elements included in the RSP that were not in the Preparedness Framework are a commitment to publicizing results of evaluation, incident reporting mechanisms, and detailed commitments for infosecurity and cybersecurity.[44]
2024 January 17 International AIGA Releases AI Guidelines International Policy At the 2024 annual World Economic Forum in Davos, the AIGA, in collaboration with IBM Consulting and Accenture, releases three reports on AI regulation and governance.[45] “Presidio AI Framework: Towards Safe Generative AI Models,” “Unlocking Value from Generative AI: Guidance for Responsible Transformation,” and “Generative AI Governance: Shaping Our Collective Global Future,” aim to address the digital divide and to apply and mobilize AI resources in sectors like healthcare and education.[46]
2024 February 8 United States Artificial Intelligence Act Organization U.S. Secretary of Commerce Gina Raimondo announces "the creation of the U.S. AI Safety Institute Consortium (AISIC), which will unite AI creators and users, academics, government and industry researchers, and civil society organizations in support of the development and deployment of safe and trustworthy artificial intelligence (AI)." AISIC is to be housed under the U.S. AI Safety Institute, and includes over 200 member organizations.[47] The member organizations were recruited through a notice published the Federal Register asking interested organizations to submit a letter of interest over a period of 75 days (between November 2, 2023, and January 15, 2024).[48][49]
2024 March 7 (anticipation), April 16 (official announcement) United States US AI Safety Institute Organization U.S. Secretary of Commerce Gina Raimondo announces additional members of the executive leadership of the U.S. AI Safety Institute (AISI); one of these is Paul Christiano as head of AI safety.[50] A month prior, when there was anticipation of this appointment VentureBeat had reported dissatisfaction with the idea of appointing Christiano, from "employees who fear that Christiano’s association with EA and longtermism could compromise the institute’s objectivity and integrity."[51][52]
2024 May 21 Europe Artificial Intelligence Act International Policy The European Council approves the Artificial Intelligence Act.[22] This Act is the first of its kind and operates within a risk-based approach - the higher the risk to society, the stricter the rules.

See also


References

  1. ↑ 1.0 1.1 "Canada is a global AI leader|". cifar.ca. 2017. Retrieved 18 September 2024. 
  2. ↑ "Pan-Canadian Artificial Intelligence Strategy|". dig.watch. June 2017. Retrieved 18 September 2024. 
  3. ↑ "China issues guideline on artificial intelligence development|". english.gov.cn. 20 July 2017. Retrieved 6 September 2024. 
  4. ↑ 4.0 4.1 "What is GDPR|". GDPR.EU. Retrieved 28 August 2024. 
  5. ↑ "The EU General Data Protection Regulation|". HRW.org. 6 June 2018. Retrieved 28 August 2024. 
  6. ↑ "California Consumer Privacy Act (CCPA)|". oag.ca.gov. Retrieved 30 August 2024. 
  7. ↑ Ocampo, Danielle (June 2024). "CCPA and the EU AI ACT". calawyers.org. Retrieved 30 August 2024. 
  8. ↑ "Maintaining American Leadership in Artificial Intelligence|". Federalregister.gov. Retrieved 30 August 2024. 
  9. ↑ Metz, Cade (11 February 2019). "Trump Signs Executive Order Promoting Artificial Intelligence". nytimes.com. Retrieved 30 August 2024. 
  10. ↑ "Human Centric AI|" (PDF). cas.co. 29 March 2019. Retrieved 15 September 2024. 
  11. ↑ Habuka, Hiroki (14 February 2023). "Japan's Approach to AI Regulation and Its Impact on the 2023 G7 Presidency". CSIS.org. Retrieved 15 September 2024. 
  12. ↑ Habuka, Hiroki (14 February 2023). "Japan's Approach to AI Regulation and Its Impact on the 2023 G7 Presidency". CSIS.org. Retrieved 15 September 2024. 
  13. ↑ "116th Congress|". Congress.gov. Retrieved 30 August 2024. 
  14. ↑ Robertson, Adi (10 April 2019). "A new bill would force companies to check their algorithms for bias". theverge.com. Retrieved 30 August 2024. 
  15. ↑ "117th Congress|". Congress.gov. Retrieved 30 August 2024. 
  16. ↑ 16.0 16.1 "OECD AI Principles overview|". oecd.ai. Retrieved 11 September 2024. 
  17. ↑ "AI Principles|". oecd.org. Retrieved 11 September 2024. 
  18. ↑ "Australia's AI Ethics Principles|". industry.gov.au. 7 November 2019. Retrieved 18 September 2024. 
  19. ↑ Tonkin, Casey (7 November 2019). "AI ethics framework being put to the test". ia.acs.org.au. Retrieved 18 September 2024. 
  20. ↑ 20.0 20.1 20.2 "About GPAI|". gpai.ai. June 2020. Retrieved 20 September 2024. 
  21. ↑ O’Brien, Matt (29 May 2020). "US joins G7 artificial intelligence group to counter China". defensenews.com. Retrieved 20 September 2024. 
  22. ↑ 22.0 22.1 22.2 22.3 "European Council - Council of the European Union|". consilium.europa.eu. Retrieved 30 August 2024. 
  23. ↑ "National AI Strategy|" (PDF). gov.uk. 22 September 2021. Retrieved 15 September 2024. 
  24. ↑ Marr, Bernard (3 November 2021). "The Future Role Of AI And The UK National AI Strategy – Insights From Professor Mark Girolami". forbes.com. Retrieved 15 September 2024. 
  25. ↑ Roman, Juliana (4 October 2021). "Artificial Intelligence in Brazil: the Brazilian Strategy for Artificial Intelligence (BSAI/EBIA) and Bill No. 21/2020". irisbh.com.br. Retrieved 18 September 2024. 
  26. ↑ Lowe, Josh (13 April 2021). "Brazil launches national AI strategy". globalgovernmentforum.com. Retrieved 18 September 2024. 
  27. ↑ Roman, Juliana (4 October 2021). "Artificial Intelligence in Brazil: the Brazilian Strategy for Artificial Intelligence (BSAI/EBIA) and Bill No. 21/2020". irisbh.com.br. Retrieved 18 September 2024. 
  28. ↑ Tedeneke, Alem (15 June 2023). "World Economic Forum Launches AI Governance Alliance Focused on Responsible Generative AI". weforum.org. Retrieved 20 September 2024. 
  29. ↑ Tedeneke, Alem (15 June 2023). "World Economic Forum Launches AI Governance Alliance Focused on Responsible Generative AI". weforum.org. Retrieved 20 September 2024. 
  30. ↑ "FACT SHEET: Biden-⁠Harris Administration Secures Voluntary Commitments from Leading Artificial Intelligence Companies to Manage the Risks Posed by AI|". whitehouse.gov. 21 July 2023. Retrieved 11 September 2024. 
  31. ↑ Chatterjee, Mohar (21 July 2023). "White House notches AI agreement with top tech firms". politico.com. Retrieved 11 September 2024. 
  32. ↑ 32.0 32.1 32.2 "Anthropic's Responsible Scaling Policy |". anthropic.com. 19 September 2023. Retrieved 11 September 2024. 
  33. ↑ "Responsible Scaling: Comparing Government Guidance and Company Policy |". iaps.ai. 11 March 2024. Retrieved 11 September 2024. 
  34. ↑ "FACT SHEET: President Biden Issues Executive Order on Safe, Secure, and Trustworthy Artificial Intelligence|". whitehouse.gov. Retrieved 30 August 2024. 
  35. ↑ "NIST Calls for Information to Support Safe, Secure and Trustworthy Development and Use of Artificial Intelligence|". nist.gov. 19 December 2023. Retrieved 30 August 2024. 
  36. ↑ Mowshowitz, Zvi (1 November 2023). "Reactions to the Executive Order". lesswrong.com. Retrieved 30 August 2024. 
  37. ↑ "AI Executive Order Timeline|". bipartisanpolicy.org. 13 December 2023. Retrieved 30 August 2024. 
  38. ↑ "The Bletchley Declaration by Countries Attending the AI Safety Summit, 1-2 November 2023". GOV.UK. November 1, 2023. Retrieved May 19, 2024. 
  39. ↑ Soares, Nate (October 31, 2023). "Thoughts on the AI Safety Summit company policy requests and responses". LessWrong. Retrieved May 19, 2024. 
  40. ↑ Mowshowitz, Zvi (November 7, 2023). "On the UK Summit". LessWrong. Retrieved May 19, 2024. 
  41. ↑ "FACT SHEET: Vice President Harris Announces New U.S. Initiatives to Advance the Safe and Responsible Use of Artificial Intelligence". White House. November 1, 2023. Retrieved July 6, 2024. 
  42. ↑ "Prime Minister launches new AI Safety Institute. World's first AI Safety Institute launched in UK, tasked with testing the safety of emerging types of AI.". GOV.UK. November 2, 2023. Retrieved July 6, 2024. 
  43. ↑ "OpenAi Preparedness Framework (Beta)|" (PDF). openai.com. 18 December 2023. Retrieved 11 September 2024. 
  44. ↑ 44.0 44.1 44.2 "Is OpenAI's Preparedness Framework better than its competitors' "Responsible Scaling Policies"? A Comparative Analysis|". safer-ai.org. 19 January 2024. Retrieved 11 September 2024. 
  45. ↑ Sibahle, Malinga (18 January 2024). "AI Governance Alliance debuts research reports on AI guidelines". itweb.co.za. Retrieved 20 September 2024. 
  46. ↑ Sibahle, Malinga (18 January 2024). "AI Governance Alliance debuts research reports on AI guidelines". itweb.co.za. Retrieved 20 September 2024. 
  47. ↑ "Biden-Harris Administration Announces First-Ever Consortium Dedicated to AI Safety". U.S. Department of Commerce. February 8, 2024. Retrieved July 6, 2024. 
  48. ↑ "Artificial Intelligence Safety Institute Consortium". Federal Register. November 2, 2023. Retrieved July 6, 2024. 
  49. ↑ "Artificial Intelligence Safety Institute Consortium (AISIC)". National Institute of Standards and Technology. Retrieved July 6, 2024. 
  50. ↑ "U.S. Commerce Secretary Gina Raimondo Announces Expansion of U.S. AI Safety Institute Leadership Team". U.S. Department of Commerce. April 16, 2024. Retrieved July 6, 2024. 
  51. ↑ Goldman, Sharon (March 7, 2024). "NIST staffers revolt against expected appointment of 'effective altruist' AI researcher to US AI Safety Institute". VentureBeat. Retrieved July 6, 2024. 
  52. ↑ "NIST staffers revolt against expected appointment of 'effective altruist' AI researcher to US AI Safety Institute (linkpost)". Effective Altruism Forum. March 8, 2024. Retrieved July 6, 2024. 
Retrieved from "http://timelines.issarice.com/index.php?title=Timeline_of_AI_policy&oldid=82146"