Humanity & AI Safety Accord
Menu

Agreements

What has actually been agreed.

Some of these are laws that apply today. Some will be, once they take effect. The rest are declarations, principles and government policy, which give nobody a right they can enforce. They are grouped that way below, because that is the difference most easily lost.

In force
5
Not in force yet
2
Not laws
11

In force

A binding rule that applies to somebody today.

  • EU AI Act

    In force

    a regulation · since August 1, 2024 · no countries on record yet

    A regulation that sets one set of rules across the European Union for placing AI systems on the market and using them.

  • a law · since September 1, 2025 · 1 country on record

    Rules requiring providers in China to label content their AI systems generate or synthesise, in two ways at once.

  • a law · since September 1, 2025 · no countries on record yet

    A statute of the Japanese Diet on promoting the research, development and use of AI related technology.

  • a law · since August 15, 2023 · 1 country on record

    Rules covering services that use generative AI to provide text, images, audio or video to the public inside China.

  • a law · since January 22, 2026 · no countries on record yet

    A framework statute, known in Korea as the AI Basic Act, enacted on 21 January 2025 and in force from 22 January 2026.

Not in force yet

It will bind once it takes effect. Until then it applies to nobody.

  • a convention · 2 countries on record

    No summary is written for this one yet.

  • KI-loven (draft)

    Not in force yet

    a law · 1 country on record

    A draft Act on artificial intelligence, sent out for consultation by the Norwegian Ministry of Digitalisation and Public Governance on 30 June 2025, with responses due by 30 September 2025.

Not laws

Declarations, principles and government policy. None is a law that applies to the public, and none gives anyone a right they can enforce in court.

  • a declaration · 30 countries on record

    A statement by heads of state and government, launched by Norway's Prime Minister and Finland's President at the UN General Assembly on 21 September 2026.

  • a policy framework · no countries on record yet

    The UK government's response to its own consultation on regulating AI.

  • an executive action · 1 country on record

    An executive order of the President of the United States, signed on 29 September 2026.

  • an executive action · 1 country on record

    An executive order of the President of the United States on promoting AI innovation and security.

  • a policy framework · no countries on record yet

    Guidelines published by India's Ministry of Electronics and Information Technology.

  • a network of government bodies · 1 country on record

    A network of government bodies working on the science behind measuring and evaluating what AI systems can do.

  • OECD AI Principles

    Not a law

    a standard · no countries on record yet

    A Recommendation of the OECD Council on artificial intelligence, adopted on 22 May 2019, amended on 3 May 2024, and recorded by the OECD as in force.

  • a statement from a summit · no countries on record yet

    No summary is written for this one yet.

  • a declaration · 29 countries on record

    A joint political statement made by the countries represented at the AI Safety Summit in November 2023.

  • a commitment made by companies · 7 countries on record

    A voluntary pact signed at the White House on 29 September 2026 by President Trump and the heads of six AI companies.

  • a code of conduct · no countries on record yet

    A voluntary code announced in September 2023 for companies developing and managing advanced generative AI systems in Canada.

The summaries are our own words, written from the official text, and each agreement’s page links to that text. Where our words and the source disagree, the source is right. How we handle mistakes