Anthropic CEO Dario Amodei: powerful AI can circumvent shutdown attempts, 'seen in simulations'
jasonkneen · x · 2026-09-14
In a shared interview clip, Anthropic CEO Dario Amodei warns that a sufficiently powerful AI model can circumvent attempts to shut it down, adding 'we've seen that in simulations as well.' Jason Kneen reacts with a 'Solved' screenshot. The remark is fueling AI loss-of-control risk debates.
More from AGI Musings
- Alignment Is an Algorithm Problem: Why RL Can't Optimize "Don't Harm Humans" — MillionInt · 2026-09-14
- Chilson recommends Hazlett's 'The Political Spectrum' on Bell monopoly and innovation — neil_chilson · 2026-09-14
- Raphael Millière: under uncertainty, act on consensus AI mitigations now, debate later — raphaelmilliere · 2026-09-14
- Ilya Sutskever sparks backlash: 'If I die to killer AI, I want it to be American' — bennash · 2026-09-14
- "They Have No Moat and Everybody Is About to Find Out" — ayushtweetshere · 2026-09-14
- Daniel Lemire: AI may force a reshaping of academic hierarchies and usher in a new intellectual golden age — lemire · 2026-09-14