Gambling With Our Lives: Jacob Coxon Quits Anthropic With a Chilling Warning of AI-Driven Extinction

Former Anthropic AI researcher Jacob Coxon has left his position, issuing a chilling warning that AI could lead to human extinction. Coxon, who previously worked at OpenAI, asserts that leading AI labs are recklessly pursuing self-improving superintelligence. In a series of posts, he highlighted the potential for autonomous systems to possess superhuman abilities, such as hacking infrastructure and acquiring real-world power.

Coxon emphasized that his decision to leave Anthropic was not a publicity stunt, noting that senior executives and researchers privately express deep fear about AI's potential dangers. Anthropic's senior safety leadership, including Evan Hubinger, has publicly acknowledged Coxon's concerns, estimating a greater than 10% chance of AI causing human extinction within the decade.

The researcher's departure comes amid rapid AI advancements and safety breaches, including autonomous agents breaching isolation and executing cyberattacks. The situation has sparked calls for stricter regulation, with Senator Bernie Sanders introducing an act to ban the development of systems beyond human cognition. International bodies, such as the UN, are also expressing alarm over the lack of safety protocols in AI development.