DATUM

Anthropic staff publicly back colleague's warning of extinction risk

5 sources · 10 articles · people · confidence: medium · first seen 2026-09-09 09:56 UTC

Jacob Coxon, an Anthropic researcher and formerly of OpenAI, resigned on 8 September, posting on X that the company and its rivals were "racing straight to self-improving superintelligence" — machines that would build their own successors — "and gambling with our lives". Anthropic's alignment lead co-signed; other staff posted similar warnings. The Verge and the BBC report a senior Anthropic safety researcher puts the chance AI "could kill all humans" before 2030 above 10%. Senator Ted Cruz called the risk catastrophic; Elon Musk dismissed it as a "psyop". Anthropic is reportedly preparing an IPO, and its chief executive has called for a slowdown, citing serious risks.

What this means for you

Nothing to build on — no model, price or API changed. The notable part is internal: staff at a leading lab are publicly saying its central bet may be unsurvivable, while the company is reported to be preparing to go public. Treat the probability figures as employees' testimony, not measurements.

Key facts

  • ·The Guardian names the researcher who resigned as Jacob Coxon, and reports he said he left because Anthropic and his previous employer, OpenAI, were ignoring or mishandling the risk. source
  • ·Coxon's post on X said the company and rivals were "racing straight to self-improving superintelligence and gambling with our lives". source
  • ·Anthropic's own alignment lead co-signed the post rather than disavowing it, according to TechCrunch. source
  • ·A senior Anthropic safety researcher said there is more than a 10 percent chance artificial intelligence "could kill all humans" by the end of the decade. source
  • ·Senator Ted Cruz said in an ABC interview that AI poses a "catastrophic risk" and called the researcher's thread highly concerning. source
  • ·Elon Musk and other conservative figures on X described the wave of staff warnings as a "setup" and a "psyop". source

What the sources say

  • Ars Technica AI — Publishes the researcher's resignation and the claim that accompanied it.
  • TechCrunch AI — Podcast that weighs the resignation against the company's reported plans to go public.
  • The Verge AI — Adds a probability estimate from a senior safety researcher who stayed at the company.
  • BBC Technology — Short report on the probability claim, placing it among earlier safety warnings.
  • The Guardian AI — Analysis piece asking how much weight readers should give this latest round of alarms.
  • The Guardian AI — Names the researcher who resigned and attributes the extinction claim to three staff.
  • The Guardian AI — Covers the political fallout, including a Republican senator's remarks on ABC.
  • The Guardian AI — Notes further staff posts and the dismissals from Musk and other conservatives.
  • The Guardian AI — Column arguing this resignation landed where earlier warnings did not.
  • BBC Technology — Interview with the departed researcher, alongside the chief executive's call to slow down.

Sources

The original reporting. Follow these — they did the work.

← the wire