A group of countries has published A Call for Control of Frontier AI Models, a joint leaders’ statement launched during the UN General Assembly week. The version published by the Government of the Netherlands lists 21 signatories, including Germany, Canada, Australia, Singapore, Spain, the United Arab Emirates, South Africa, and the European Commission. It is a political statement, not binding law.
Key points
- Its core principle is that AI must remain under human direction, oversight, and control, in line with international law.
- It calls on developers to adopt transparent safety protocols, including mandatory pre-deployment testing and independent evaluation.
- It asks governments to develop and coordinate common standards, strengthen transparency, and build access to trusted evaluation capacity.
- It asks UN member states to explore an international institution that could set standards and verify compliance.
The statement does not create obligations, but it signals the direction of travel for frontier AI oversight in many of the countries that signed it. Organisations building on frontier models should expect more questions about testing, evaluation, and incident reporting, which an ISO/IEC 42001 management system or the NIST AI RMF can help to answer.
Source: Government of the Netherlands
This update is general information, not legal advice. If you would like help assessing the impact on your information security or privacy programme, contact us.
