Cambridge safety researcher David Krueger: Anthropic's plan is too little, too late

DavidSKrueger · x · 2026-09-13

Cambridge AI safety researcher David Krueger criticizes Dario Amodei's recently announced safety plan as "too little, too late": it does not reduce risk to an acceptable level, and Dario carefully avoids claiming it would. He asks why we should settle for it when better plans might exist, thanks Rep. Don Beyer for voicing similar criticism, and links a thread outlining plans for how to stop.

Related event: Cambridge Researcher: Dario's Safety Plan Too Little, Too Late(3 posts)→

Original post →

More from Safety

Safety channel →