At a moment when artificial intelligence reshapes the foundations of work, knowledge, and power, one of the industry's most safety-conscious voices has acknowledged what critics long argued: that companies cannot be trusted to govern themselves. Anthropic's policy leadership has called for binding external oversight, not as a concession, but as a recognition that competitive pressure and misaligned incentives make voluntary compliance structurally untenable. The declaration reframes the question of AI governance — not as a debate between innovation and restriction, but as a reckoning with who