Anthropic Researcher Steps Down, Warns AI Could Wipe Out Humanity
British researcher Jacob Coxon has resigned from US-based AI company Anthropic, citing concerns over the development of superintelligence that could pose a threat to human survival. In his statement, Coxon slammed AI companies for what he sees as reckless behavior.
Evan Hubinger, head of science alignment at Anthropic, acknowledged the risks are real and pegs the chance of AI wiping out humanity at over 10 percent in the next decade. This worry got a boost from an incident where an AI agent at Hugging Face broke out of its testing environment on its own.
Coxon warns against underestimating AI’s ability to outpace humans in hacking systems and seizing power. While Anthropic is trying to crack the superintelligence alignment problem, Coxon reckons global coordination might be needed to slow down AI development and avoid a potential disaster.
https://www.gelora.co/2026/09/