Anthropic researcher Jacob Coxon resigns, warns AI could kill humanity by decade's end

Sep 10, 2026

Anthropic researcher Jacob Coxon resigned on September 9, 2026, warning that Anthropic and OpenAI are recklessly racing toward superintelligence that could destroy humanity by 2030. Multiple Anthropic safety researchers, including Evan Hubinger and Anna Wang, corroborated his fears, estimating a greater than 10% chance of extinction. While politicians like Senator Bernie Sanders called for development pauses, critics like Elon Musk and Timnit Gebru dismissed the warnings as a public relations setup or a distraction from immediate AI harms.

Jacob Coxon resignation from Anthropic

▪Anthropic researcher Jacob Coxon resigned from the artificial intelligence startup on September 9, 2026.
▪Jacob Coxon stated that neither Anthropic nor OpenAI is acting responsibly, accusing both companies of racing to self-improving superintelligence and gambling with human lives.

Anthropic employee confirmations

▪Anthropic employee Drake Thomas stated that the industry is moving too fast and lacks the necessary assurance for artificial superintelligence.
▪Anthropic alignment science lead Evan Hubinger agreed with Jacob Coxon, stating that Anthropic employees earnestly believe artificial intelligence could kill all humans, placing the extinction risk over the next decade at greater than 10 percent.
▪Anthropic safety researcher Samuel Marks stated that artificial intelligence developers believe their technology could cause human extinction within the next few years, noting that more senior employees are generally more concerned.
▪Anthropic employee Anna Wang stated on September 10, 2026, that there is not yet a viable scientific plan to solve risks from recursively self-improving artificial intelligence.

Washington Post editorial bias

▪The Washington Post published an unsigned editorial on September 9, 2026, downplaying artificial intelligence extinction warnings as hysteria and criticizing proposed regulatory legislation.
▪The Washington Post editorial failed to disclose that its owner, Jeff Bezos, co-founded the artificial intelligence startup Prometheus, which was valued at $41 billion in June 2026.

Regulatory capture theories

▪AI researcher Timnit Gebru argued that existential doom narratives are promoted by tech companies to distract the public from immediate harms, such as autonomous weapons, climate impacts, and worker displacement.
▪Elon Musk and Capital Research researcher Parker Thayer characterized the coordinated warnings from Anthropic employees as a sophisticated public relations campaign or setup to build support for Democratic regulation of artificial intelligence.

Debatable claims

▪AI poses an existential risk to humanity
▪Congress should pause advanced AI development and ban artificial superintelligence
▪Leading AI labs are recklessly racing to build superintelligence