Safety, risk management, transparency, and liability for high-risk and frontier systems. A withdrawn liability text is labeled as an idea. Also here: the Council of Europe’s treaty on artificial intelligence, human rights, democracy and the rule of law; the ethics recommendation of the United Nations Educational, Scientific and Cultural Organization (UNESCO); the AI principles of the Organisation for Economic Co-operation and Development (OECD); the AI Risk Management Framework of the U.S. National Institute of Standards and Technology; and the United Kingdom’s approach of regulating AI through its existing regulators.
Texts that are in force: a statute, regulation, directive, or treaty that already binds someone.
Regulation (EU) 2024/1689 of 13 June 2024 laying down harmonised rules on artificial intelligence
The first horizontal EU law that sorts AI systems by risk. High-risk duties were later deferred; some transparency rules kept their original date.
Read the article
Act No. 20676, 21 January 2025 (as later amended); commonly called the AI Basic Act
A national AI statute in force from January 2026, with a stated grace period on many fines.
Read the article
Cyberspace Administration of China and six other departments; published 13 July 2023; effective 15 August 2023
Binding rules for public generative-AI services inside China: filing, content, and training-data duties for those public services.
Read the article
Senate Bill 53 (Wiener), Chapter 138, Statutes of 2025; California Business and Professions Code sections 22757.10 and following
A state statute for the largest model developers: publish a safety framework and report serious incidents. Senate Bill 1047 of 2024 was vetoed.
Read the article
Adopted by the General Conference of UNESCO in November 2021
The first global standard on the ethics of artificial intelligence, adopted by all 193 Member States of UNESCO. Human rights and human dignity sit at its centre.
Read the article
Adopted 22 May 2019; revised 8 November 2023 and 3 May 2024
The first intergovernmental standard on artificial intelligence, adopted in May 2019 and updated in May 2024. On 29 September 2026, 50 countries and the European Union had committed to it.
Read the article
Version 1.0, released 26 January 2023
A voluntary guide from the U.S. National Institute of Standards and Technology that helps organisations identify, measure and manage the risks of AI systems. Released in January 2023 and now being revised.
Read the article
White paper, 29 March 2023; government response, 6 February 2024
The UK government set five principles for existing regulators to apply in their own sectors, instead of passing a single new AI law. Its February 2024 response kept that approach and added funding and deadlines for regulators.
Read the articleBills in a legislature and formal intergovernmental negotiations. A first reading, and a round of an intergovernmental negotiating committee, are stages on the way to a later decision.
Bill C-36, 45th Parliament, 1st session: an Act to enact the Protecting Privacy and Consumer Data Act, to amend the Personal Information Protection and Electronic Documents Act, and to make related amendments
A live privacy bill with automated-decision duties. It is a separate text from the Artificial Intelligence and Data Act, and it is still at first reading.
Read the article
Council of Europe Treaty Series No. 225, opened for signature on 5 September 2024
The first international treaty on artificial intelligence that becomes binding on the countries that ratify it. On 29 September 2026 it had 21 signatories and one ratification, by the European Union.
Read the articleUseful proposed laws that remain proposals. Some were formally filed and then withdrawn or left to lapse. Read them as proposals.
