Anthropic staffer: >10% chance AI kills all humans within a decade, no alignment plan yet
StewartalsopIII · x · 2026-09-10
An Anthropic employee (EvanHub) said the team earnestly believes AI could kill all humans, personally estimating >10% probability within the next decade. He acknowledged Anthropic is trying its best but admits there is not yet a plan to solve alignment for superintelligence, nor a clear path to one. The remark circulated widely via a retweet by 0xDesigner.
More from AGI Musings
- Anthropic researcher puts >10% odds on AI killing all humans within a decade, says alignment unsolved — StewartalsopIII · 2026-09-10
- The AI doom playbook: from speculative assumptions to public panic in five steps — DrTechlash · 2026-09-10
- AI doomerism decoded: rationalist mythos, MIRI's tone-setting and EA funding infrastructure — DrTechlash · 2026-09-10
- Why the 'Doom Bible' fails to convince: anthropomorphism, compressed timelines and ungrounded scenarios — DrTechlash · 2026-09-10
- AI panic messaging is deliberately A/B tested, ten studies on media fear cycles find — DrTechlash · 2026-09-10
- A thread debunking AI doomerism: weak foundations, panic messaging and subculture origins — DrTechlash · 2026-09-10