22 countries have backed a declaration calling for artificial intelligence to remain under human direction, oversight and control, while urging the development of stronger international safeguards for advanced AI systems. Adopted on the sidelines of the United Nations General Assembly, the declaration recognises AI’s potential to improve lives, advance science and boost economies, but warns that the rapid development of frontier models could create serious safety and security risks if not properly managed.
The initiative was led by Finnish President Alexander Stubb and Norwegian Prime Minister Jonas Gahr Støre, with Stubb telling POLITICO he hopes to build international support, including from the United States and China. The declaration calls for common safety standards, greater transparency, mandatory pre-deployment testing and independent evaluations of advanced AI systems.
It also proposes exploring the creation of an international institution to oversee AI, potentially modelled on the International Atomic Energy Agency. The initiative does not call for slowing or pausing AI development, instead focusing on managing risks while allowing the technology’s potential benefits to be realised.
via Politico
