Axios reports on a new essay from the AI Futures Project, the group behind the ‘AI 2027’ forecast, calling for deliberately slowing the race toward superintelligence until around 2040. The six authors, including former OpenAI researcher Daniel Kokotajlo and AI 2027 co-author Thomas Larsen, recommend an international agreement among major AI powers to scale capabilities slowly and safely rather than race in secrecy. They propose AI companies disclose everything except model weights so outside groups can verify safety claims.
