OpenAI argues that coordinated action between individual states and the federal government is essential to establish a coherent national standard for frontier AI safety. The company highlights California, New York, and Illinois as examples of states aligning on core safeguards through "reverse federalism," creating a de facto baseline while federal frameworks are developed.
- States should adopt documented safety frameworks with risk assessments, incident reporting, and independent audits to prevent regulatory patchworks.
- The federal government must lead testing and evaluation of advanced models through the Center for AI Standards and Innovation (CAISI).
- Developers of capable systems must meet requirements including independent audits, security standards, and whistleblower protections.
- Federal legislation should support a US-led international forum to establish global AI standards based on democratic values.
This approach aims to ensure democratic oversight of AI deployment while maintaining US leadership in innovation and preventing chaotic or inconsistent regulations.