Jacob Coxon, a researcher who trained AI systems at Anthropic, announced his resignation from the company, citing concerns about its approach to artificial intelligence safety. In his departure statement, Coxon accused both Anthropic and OpenAI of moving too rapidly toward developing self-improving superintelligent systems without adequate safeguards, despite acknowledging the existential risks involved.
Coxon characterized the situation as a competitive race between AI companies to achieve advanced capabilities first, with both firms prioritizing speed over careful risk mitigation. He noted that researchers at these organizations genuinely believe advanced AI could pose an existential threat to humanity within the next decade, yet companies continue accelerating development timelines.
In response to Coxon’s departure, Evan Hubinger, who leads an AI safety team at Anthropic, publicly confirmed these concerns. Hubinger stated that self-improving AI systems are developing faster than previously anticipated and that he personally estimates a greater-than-10-percent probability of AI causing human extinction within the next decade. He further disclosed that Anthropic currently lacks a comprehensive safety plan for ensuring advanced AI systems remain controlled and aligned with human values, and the company is not clearly making sufficient progress toward developing one.
Coxon’s exit represents a significant personnel loss for Anthropic, which was itself founded by former OpenAI executives concerned about safety practices at that organization. His departure joins a pattern of researchers citing safety concerns when leaving major AI companies. The situation reflects growing unease within the AI research community regarding the pace of development, the competitive pressures facing major labs, and the potential risks of rapidly advancing AI capabilities without corresponding safety breakthroughs.
Article Attribution | Read More at Article Source
Article summary produced by Claude AI