U.S. and China Will Start Discussing A.I. Safety, Bessent Says
The New York Times • • nation
Key Points:
- The United States and China plan to discuss establishing guardrails on artificial intelligence, including protocols to prevent powerful AI models from falling into the hands of nonstate actors, according to U.S. Treasury Secretary Scott Bessent.
- These talks, if they occur, would mark the first formal discussions on AI between the two countries during President Trump's second term, amid growing concerns about AI being weaponized or losing human control.
- Bessent noted that the U.S. is willing to cooperate with China on AI safety because China is currently behind the U.S. in AI development, allowing the U.S. to set best practices and values for global adoption.
- Experts estimate that Chinese AI models lag a few months behind leading U.S. models, which influences the dynamics of potential cooperation.
- A challenge to U.S.-China collaboration on AI safety is their differing focus on potential AI threats, complicating unified approaches.