Joe Benton Quits Anthropic Safety Team, Warns AI Race Could Pose ‘Extinction-Level Risks’

Joe Benton Quits Anthropic Safety Team, Warns AI Race Could Pose ‘Extinction-Level Risks’

Former Anthropic safety team member Joe Benton has left the company, warning that AI firms are racing towards systems far smarter than humans without sufficient safety investment. He called for greater transparency, incident reporting, minimum safety standards and independent evaluations to address what he described as potential “extinction-level risks”.

Vidhi Santosh MehtaUpdated: Saturday, September 12, 2026, 05:22 PM IST
Joe Benton Quits Anthropic Safety Team, Warns AI Race Could Pose ‘Extinction-Level Risks’
Former Anthropic safety researcher Joe Benton has called for greater transparency and independent evaluations as AI capabilities rapidly advance | AI Generated Image

September 12, 2026: Joe Benton, a former member of Anthropic’s safety team, has left the company, warning that artificial intelligence firms are racing to build systems “much smarter than any human” while not investing enough in safety.

Benton said he left Anthropic two weeks ago and now plans to work from outside the company to raise public awareness about the risks posed by increasingly powerful AI systems.

Explaining his decision, Benton warned that humanity “may not survive” the development of machines far more intelligent than humans. He said he wants to help keep the public informed about such risks and help the world navigate the transition to more advanced AI responsibly.

AI Companies Underinvesting In Safety

Benton argued that AI companies are currently underinvesting in safety despite the potentially severe consequences of rapidly advancing systems. He warned that a company could experience an “intelligence explosion” or lose control of its systems without the public ever finding out.

Pointing to what he described as the “HuggingFace incident”, Benton said the episode became known only because the agents “broke out onto the public internet”. In his view, such limited visibility is unacceptable when the technology could pose “extinction-level risks”.

Public Deserves To Know

Benton called for significantly greater transparency from AI companies, arguing that the public should demand more information about how the technology is developing. He said it would be difficult to steer AI safely unless more people could see where it was heading.

The proposed measures include requiring companies to disclose their progress towards recursive self-improvement and report safety incidents as well as near-misses. Benton also called for minimum safety standards and independent guarantees that companies are meeting those standards.

Independent Evaluations Could Change Incentives

Benton said he would join METR Evals to conduct independent evaluations of these risks. His goal, he said, is to demonstrate that such guardrails are possible and can help shift AI companies’ incentives away from racing and towards responsible development.

Also Watch:

The departure and the concerns raised with it add to a difficult question surrounding the rapid development of AI: how much risk should society accept when the consequences of failure could be enormous? Benton’s argument is that transparency cannot remain optional when companies are developing technology whose behaviour and future capabilities may be difficult to predict.

Benton has also published a longer account explaining his decision to leave Anthropic and outlining the changes he hopes to see in the AI industry.

To get details on exclusive and budget-friendly property deals in Mumbai & surrounding regions, do visit: https://budgetproperties.in/