Response to NIST Executive Order on AI

OpenAI responded to NIST’s AI executive-order request with guidance on evaluating harmful model capabilities and strengthening safeguards.

The company highlighted the use of domain experts, red teaming, and its Preparedness Framework to assess risks before deployment.

OpenAI also pointed to work on synthetic media provenance, model autonomy, and CBRN-related dangers as key safety priorities.

The response could influence federal standards for testing, auditing, and governing advanced AI systems.