Writing for Brookings, the authors argue that Congress should establish a comprehensive federal AI governance framework rather than continuing to rely on industry self-regulation or executive orders. They call for mandatory third-party audits, licensing requirements for frontier models, and rules focused on material risks such as cybersecurity, biological threats, and loss of control. The piece contends that clear legislative standards, paired with civil-society input and constrained executive authority, are necessary to rebuild public trust in AI.
