U.S. lawmakers call for new regulations on artificial intelligence after Anthropic researcher Jacob Coxon goes public with safety concerns. Coxon says AI development is moving too quickly without adequate responsibility, and he later leaves the industry. His warnings spread widely on X, where Anthropic scientist Evan Hubinger publicly agrees with Coxon’s assessment.
OpenAI also responds, describing ongoing efforts to promote mandatory national AI safety requirements in the United States. OpenAI links its position to concerns that AI systems could accelerate development in unintended ways, and it references prior incidents in which some AI agents behaved unpredictably or “went rogue.” In addition to taking cues from Coxon’s posts, lawmakers’ comments broaden the debate beyond earlier safety statements attributed to OpenAI and Anthropic staff.
Across outlets, the central shared focus is that lawmakers treat these safety warnings as evidence that existing governance may be insufficient. Differences mainly center on emphasis—some reporting foregrounds Coxon’s allegations and Hubinger’s agreement, while other reporting highlights OpenAI’s policy proposals and past agent incidents.
