Global / AI Safety
Anthropic researcher leaves, citing existential AI risk
Safety concerns about superintelligence are driving talent away from leading labs.
Researcher Jacob Coxon has left Anthropic and the AI industry, warning that companies are moving too quickly toward superintelligence. Anthropic's own Evan Hubinger estimates a 10% chance of AI killing all humans within the next decade.
Published · significance 57 of 100 (medium) · 1 source
What happened
Jacob Coxon, an AI researcher at Anthropic, has quit the company and exited the AI industry entirely. He cited concerns about the pace of development toward superintelligence at companies like Anthropic and OpenAI. Evan Hubinger, another Anthropic researcher, has publicly estimated a greater than 10% probability of advanced AI posing an existential risk to humanity within the next decade.
Why it matters
This departure signals friction between safety researchers and the commercial imperative at leading labs. When researchers at frontier companies publicly assign double-digit existential risk probabilities, it reveals genuine internal disagreement on the pace of development. Talent departures driven by safety concerns undermine the recruiting advantage that frontier labs rely on.
What changes
Anthropic's safety credibility faces scrutiny when its own researchers cite existential risk while the company races toward larger models. The incident may embolden other researchers to voice similar concerns, or accelerate departures from safety-focused roles.
Involved
Sources
Written by AI from the reports above; scored by a published formula. How we work. Found a mistake? Email lockedinshreyash@gmail.com. Up To Date summarises and links to original reporting; it never reproduces articles.
Related coverage
- Anthropic researcher exits over self-improving AI safety concerns — The departure signals tension within a major safety-focused lab over how to manage systems that could improve themselves. (2026-09-09)
- Anthropic disrupts Russian and Chinese efforts to misuse Claude for cyberattacks and bioweapons research — The frontier lab is now publishing details of state-sponsored and criminal abuse attempts against its own models. (2026-09-10)
- Anthropic CEO Dario Amodei urges slowdown in frontier model development — A leading AI lab's leader warns that speed now risks systemic danger, forcing the industry to reckon with its own momentum. (2026-09-14)
- Anthropic and OpenAI to embed independent safety evaluators in their labs — Companies seek outside oversight to manage risks, but researchers doubt independence without regulatory teeth. (2026-09-16)
- Anthropic researcher Jacob Coxon resigns, citing urgent need for AI safety measures — A senior scientist's departure signals deepening friction between safety ambitions and commercial pressure inside leading labs. (2026-09-09)