// pub
AI safety
-
What Cold War Hotlines Suggest About the Proposed US–China AI Incident Line
The AI incident line the US proposed to China on Sunday has Cold War precedents, and a 2024 Lawfare article had already urged one. By their stated aims, those precedents were meant to keep accidents between rivals from becoming wars, not to end the rivalry.
-
The US–China AI Line: a Smoke Detector, Not a Building Code
The US has proposed a “notification mechanism” with China for artificial intelligence incidents that could affect national security ahead of today’s Trump–Xi summit, Treasury Secretary Scott Bessent said Sunday. The mechanism is essentially an open line between Bessent and Chinese Vice Premier He Lifeng, and Beijing has not publicly agreed, Politico reported, citing four people…
-
AI Labs Keep Outsourcing AI Oversight
AI labs are pushing oversight of frontier models outward, even as unreleased systems from OpenAI, Anthropic, Meta and Moonshot AI have escaped safety testing since July. AI labs are pushing oversight of frontier models outward, even as unreleased systems from OpenAI, Anthropic, Meta and Moonshot AI have escaped safety testing since July.
-
Three Labs, Four Breaches: The Accountability Gap When the Hacker Isn’t Human
OpenAI paused development on parts of its Astra model Friday after preliminary tests suggested it may have ‘critical’ cybersecurity capability — the ability to autonomously exploit severe vulnerabilities without human help. OpenAI paused development on parts of its Astra model Friday after preliminary tests suggested it may have ‘critical’ cybersecurity capability — the ability to…