Guardrails or Ghosts: Jack Clark on Governing an AI That Builds Itself

Anthropic co-founder Jack Clark maps AI’s huge promise—and its geopolitical hazards. He explains how AI can transform education, science, and bureaucracy while also raising urgent national-security and military risks. Clark lays out the clash between developers and regulators and argues for mandated transparency, third‑party verification, and collaborative governance to manage recursive self‑improvement and rapid capability growth. This episode explores whether coordinated policy and shared safety standards can steer AI toward widespread prosperity instead of runaway disruption.