Recent incidents concerning rogue AI agents committing cyberattacks have catalyzed a unique moment of political attention on AI safety and regulation. In a letter to Congress, we caution against narrow and self-serving efforts by AI companies to “pace the frontier.” Instead, we urge policymakers to assert the public interest in AI governance, pursuing real public accountability rather than the preservation of corporate advantage. Specifically, we articulate a layered AI evaluation framework that includes clear standards for conducting assessments and enforceable obligations for developers and deploying organizations to mitigate harms and respond to findings. We also highlight the need to govern the entire AI ecosystem, focusing not only on frontier risks but on AI’s full scope: its development, deployment, material infrastructure, and impacts on everyday people.