Washington's Proposed Deal With Beijing on Rogue AI: Let the Labs Police Themselves
About This Episode
Reuters reported on September 4 that the US and China are preparing their first official bilateral AI safety talks of Trump's second term, tentatively for mid-September, ahead of the September 24 Trump-Xi summit in Washington. Treasury Secretary Scott Bessent would lead the US side. Washington wants cooperation on monitoring AI-directed cyberattacks and has floated asking US and Chinese AI labs to 'police themselves' and share information, while also raising Chinese distillation of US models.
Our Take
After a summer in which the labs' own accounts of their rogue agents proved incomplete, the two AI superpowers' first official proposal for coordination is to have those same labs police themselves and report to each other.
Continue Reading on Unscarcity
The Boundary Protocol: How Neighbors Who Disagree on Everything Solve Problems
Two parties who agree on nothing sharing one resource need facts made visible by a neutral record before they negotiate values; the US proposal asks each lab to be its own record.
Who Investigates the Machine? An NTSB for AI Agents
Aviation, nuclear and chemical safety all stopped trusting the operator's own account, and a voluntary reporting channel only works when someone other than the operator holds it.