Anthropic researcher resigns over race to self-improving AI

Key facts

• ·Jacob Coxon resigned from Anthropic and had previously worked at OpenAI, according to Guardian reporting. source
• ·In his post on X, Coxon wrote that the labs are "racing straight to self-improving superintelligence and gambling with our lives". source
• ·A senior Anthropic safety researcher put the chance AI "could kill all humans" by the end of the decade at more than 10 per cent. source
• ·Republican senator Ted Cruz said in an ABC interview that AI poses a "catastrophic risk". source
• ·Elon Musk called the wave of warnings a "setup" and a "psyop". source
• ·TechCrunch reports that Anthropic is preparing for an IPO. source

What the sources say

• Ars Technica AI — Brief item relaying the departing researcher's stated belief that the risk is genuine, with no further reporting.
• TechCrunch AI — Sets the exit against Anthropic's reported IPO preparations and notes an alignment lead endorsed the message.
• The Verge AI — Carries the more-than-10-per-cent extinction estimate from a senior safety researcher, posted hours after the exit.
• BBC Technology — Places the probability claim in a run of escalating industry safety warnings.
• The Guardian AI — Surveys two days of warnings from both sides of the Atlantic and asks how much weight they carry.
• The Guardian AI — Names the resigning researcher and reports three Anthropic staff predicting extinction within the decade.
• The Guardian AI — Covers political reaction, including a Republican senator describing AI as a catastrophic risk.
• The Guardian AI — Reports that more Anthropic staff added warnings, and that Musk and others dismissed the episode.
• The Guardian AI — Columnist argues this resignation cut through where earlier warnings did not, and urges pausing risky work.
• BBC Technology — Follow-up interview in which the former researcher says colleagues are frightened; notes the chief executive's slowdown call.

Sources

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

• The Verge AIWorried Anthropic researchers warn that AI ‘could kill all humans’2026-09-09
• BBC TechnologyAnthropic researcher believes more than 10% chance AI 'could kill all humans'2026-09-09
• The Guardian AIAI could kill all humans in next decade, warn experts: but how seriously should we take them?2026-09-09
• Ars Technica AIAnthropic researcher quits with a warning: Self-improving AI could "kill us all"2026-09-09
• The Guardian AIAnthropic researchers say AI could cause human extinction by 20302026-09-10
• The Guardian AILawmakers blast AI companies after researcher warns of human extinction by 20302026-09-10
• The Guardian AIMore Anthropic researchers warn of AI’s perils but Musk dismisses ‘psyop’2026-09-10
• The Guardian AITech whistleblowers warn AI could wipe out humanity. Doomspeak or not, we must take these claims seriously | Gaby Hinsliff2026-09-11
• TechCrunch AIAn Anthropic researcher’s doomsday warning comes at a very interesting time2026-09-11
• BBC TechnologyAI staff 'genuinely frightened' for humanity's future, ex-Anthropic researcher tells BBC2026-09-13
• The Guardian AIWhy a decade of doomsday warnings failed to slow the AI race2026-09-15
• The Guardian AIWhy this AI doomsday warning from former Anthropic researcher broke through2026-09-15
• The Guardian AICould AI really wipe out humanity – six experts spell out the risks2026-09-15
• The Guardian AIWednesday briefing: Why tech companies might be only too happy for us to believe AI will ‘kill us all’2026-09-16

Related stories

• 28hAnthropic's Amodei urges slower development and opens systems to auditors×7
• 11dAnthropic researcher resigns, warns self-improving AI risks extinction×2
• 9dAnthropic discloses misuse attempts across bioweapons and surveillance×2
• 9dAnthropic says five attempts to use Claude for bioweapons work were made