Key facts
- The US has proposed an AI safety notification mechanism to China.
- Treasury Secretary Scott Bessent announced the proposal.
- The mechanism aims to create an open line for sharing alerts on AI system threats.
- China has not publicly acknowledged the proposed alert system.
- Experts express skepticism about the effectiveness of the proposed superficial plan.
- US Representative Ro Khanna called for a ban on recursive self-improving AI.
The Trump administration has proposed establishing an AI safety notification mechanism with China, aiming to create an open channel for sharing alerts on potential threats posed by AI systems. Treasury Secretary Scott Bessent announced the initiative, suggesting it could facilitate the sharing of "vision of common goals" between the two AI leaders.
However, the proposal faces skepticism due to ongoing US-China tensions, including increased export controls on technology and accusations of intellectual property theft. Experts like Sam Bresnick, a research fellow at Georgetown University's Center for Security and Emerging Technology, noted that the plan lacks technical experts to assess risks, calling it "BS." Su Lian Jye, chief analyst at Omdia, suggested that while the alerts might promote transparency, they are unlikely to resolve fundamental differences on AI development and competition.
US Representative Ro Khanna criticized the proposal as superficial, advocating for a more comprehensive international agreement that would ban recursive self-improving AI and establish safeguards against its misuse for weapons. Alvin Wang Graylin, a technology strategist, viewed the development as positive, noting that the scheduled follow-up meeting in November indicates mutual interest in discussion, despite differing views on AI risk assessment.
