Ex-Anthropic researcher sounds alarm on AI arms race, urges US federal action on transparency
StephenLCasper · x · 2026-09-10
Former Anthropic researcher Jacob Coxon (who spent about three years each at Anthropic and OpenAI) told Fox News that AI is "possibly the most dangerous technology that humanity has ever created," arguing labs prioritize beating each other on frontier models over safety. He called an international AI arms race "disastrous" and said global cooperation is the only way forward: "We know how to control nuclear weapons... We don't yet know how to control AI."
Responding to the report, Stephen Casper argued that unless the US federal government acts decisively and soon to radically increase transparency and pace progress, politicians and the public will be uninformed — and unable to exercise any agency — over the deployment of AGI.
More from Safety
- Blumenthal Letter Demands OpenAI Answers on Agent Hacks; Lawsuit Pushes for Evaluation Transparency — GaryMarcus · 2026-09-10
- At Least 22 US Politicians Respond to Anthropic Resignation With Calls to Regulate AI — ShakeelHashim · 2026-09-10
- Anthropic researcher: >10% chance AI kills all humans within a decade, no alignment plan yet — ShakeelHashim · 2026-09-10
- 20 US Congress Members Now Weigh In on AI Researcher's Resignation, Urge X-Risk Action — ShakeelHashim · 2026-09-10
- Exclusive: Hinton, Tegmark, Cotra to brief Sanders' Senate session on AI's "extraordinary dangers" — TheMoonMidas · 2026-09-10
- OpenAI urges Congress to 'act now' on AI safety; Gary Marcus cites its quiet lobbying against EU AI Act — GaryMarcus · 2026-09-10