GoKawiilAnthropic researcher quits, colleague admits 10% chance AI 'kills all humans' by 2030
Jacob Coxon, a researcher who previously trained systems at OpenAI and Anthropic, resigned publicly, accusing both companies of recklessly racing toward self-improving superintelligent AI without adequate safety controls. Evan Hubinger, who leads a safety team at Anthropic, responded by agreeing with Coxon's warning, estimating more than a 10 percent chance that AI could kill all humans within the next decade, while admitting Anthropic lacks a concrete plan to ensure such systems remain safe.
GoKawiil's interpretation of the reporting above, not reported fact.
These are not outside critics but insiders at one of the industry's most safety-focused labs publicly conceding they cannot control the technology they're building, which undercuts reassurances that AI companies are managing existential risks responsibly. The admission that recursive self-improvement is progressing faster than expected, combined with no clear safety roadmap, raises pressure for external regulation rather than self-governance.
- Anthropic researcher Jacob Coxon resigned, alleging AI labs are racing recklessly toward superintelligence.
- Safety team lead Evan Hubinger estimated over 10% odds AI could kill all humans within a decade.
- Anthropic reportedly has no concrete plan yet to guarantee safety of advanced self-improving AI.
Nick Bostrom "Superintelligence" (book) — If Anthropic researchers debating extinction risk has you curious, Bostrom's Superintelligence is the book that framed much of this debate about recursive self-improvement and control problems. It's a dense but readable primer that explains exactly the scenarios these AI safety researchers keep referencing.
See Nick Bostrom "Superintelligence" (book) on Amazon → Affiliate link — we may earn a commission on purchases, at no extra cost to you. Product picked by AI based on this article; it is not a tested recommendation.Source: theverge.com — Robert Hart, 2026-09-09
Published there as: “Worried Anthropic researchers warn that AI ‘could kill all humans’”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.