The "5% extinction risk" headlines were wrong — and the misframing went beyond journalists
jessi_cata · x · 2026-10-11
- jessicata points out the headline of the 2024 survey coverage ("AI researchers think there's a 5% chance of human extinction") is itself wrong, and that wrong framing was replicated across news outlets with caveats buried.
- Context: the survey covered 2,700+ AI researchers, 58% of whom saw catastrophic outcomes as non-negligible.
- As the thread clarifies, the survey actually asked about extinction or similarly permanent and severe disempowerment — so the misframing didn't originate solely with journalists.
Related event: Viral "5% extinction risk" AI expert survey figure mislabeled, critics say(8 posts)→
More from Safety
- laya-guard: local guardrail for coding agents blocks 25/27 AgentDojo attacks at ~90ms — Extreme_Ad5709 · 2026-10-11
- NIST frameworks already exist — the real question is whether AI companies implement them — AlexTensor · 2026-10-11
- Fiction Writers Complain Claude's Safety Filters Block Even Innocuous Scenes Like a Wedding Kiss — Brw_ser · 2026-10-11
- AI Audit Found an Infinite-Mint Bug in XRP's $94B Codebase, Earning a $250K Bounty — julianweisser · 2026-10-11
- Danish CPR Breach Exposed 8.8M Records; Compromised Firm Used '123456' as Admin Password — jonerp · 2026-10-11
- TansuYegen: Voluntary AI Safety Commitments Aren't Enough as Systems Act Autonomously — TansuYegen · 2026-10-11