Humanity & AI Safety Accord
Menu

Agreements

Call for Control of Frontier AI Models

A Call for Control of Frontier AI Models

Not a law

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.

What it is
a declaration
Does it bind anyone?
No. It is a political commitment, and nobody can be taken to court over it.
In force
Nothing to come into force
Opened for signature
September 21, 2026

In short

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. It says AI must remain under human direction, oversight and control, and calls on companies to adopt transparent safety protocols with mandatory testing before release and independent evaluation; on governments to coordinate common standards and share reports of serious safety incidents; and on UN member states to explore an international institution that could set standards, enable verification and convene states when capability thresholds are crossed. It is a political statement, not a treaty, and it remains open for other leaders to endorse.

Read the text this was written from

Our own words, not the instrument’s, and not a legal description. Where this and the source disagree, the source is right.

Who has taken part

This is the full list, as the official record read on October 6, 2026. A country that is not below is not on that record either. That says what the record says, and nothing about what the country intended.

Verified participation in A Call for Control of Frontier AI Models
WhoWhat they didDate
Australiaendorsed2026-09-21
Austriaendorsed2026-09-21
Bahrainendorsed2026-09-21
Canadaendorsed2026-09-21
Croatiaendorsed2026-09-23
Denmarkendorsed2026-09-21
Estoniaendorsed2026-09-21
European Unionendorsed2026-09-21
Finlandendorsed2026-09-21
Franceendorsed2026-09-23
Germanyendorsed2026-09-21
Greeceendorsed2026-09-28
Icelandendorsed2026-09-21
Irelandendorsed2026-09-21
Kazakhstanendorsed2026-09-21
Kenyaendorsed2026-09-21
Latviaendorsed2026-09-21
Liechtensteinendorsed2026-09-22
Luxembourgendorsed2026-09-22
Netherlandsendorsed2026-09-21
Norwayendorsed2026-09-21
Portugalendorsed2026-09-24
Republic of Moldovaendorsed2026-09-21
Romaniaendorsed2026-09-21
Sierra Leoneendorsed2026-09-23
Singaporeendorsed2026-09-21
South Africaendorsed2026-09-21
Spainendorsed2026-09-21
Türkiyeendorsed2026-09-21
United Arab Emiratesendorsed2026-09-21

Where to read it

Something here is wrong? Tell us. Send us a source we can open, and a person will read it.

Call for Control of Frontier AI Models, Humanity & AI Safety Accord