OpenAI asks Congress for mandatory national AI safety rules
OpenAI asked the US Congress to replace voluntary commitments with mandatory, capability-based national rules for the most advanced AI systems. Its proposal calls for testing standards, independent assessments, cybersecurity protections and incident reporting, while the company says it will keep backing state legislation until lawmakers create a federal framework.
Artificial Intelligence··Evening
The request shifts from pledges to binding rules
OpenAI asked the US Congress to establish mandatory national safety requirements for the most capable AI systems. Chief Global Affairs Officer Chris Lehane argued that voluntary commitments are no longer sufficient as systems gain the ability to accelerate AI development itself. The Next Web independently reported the request and its federal focus. OpenAI wants the rules to follow model capability, allowing the framework to change as the technology advances rather than attaching obligations to every system in the same way.[1], [2]
Testing and incident reporting form the core
The proposed framework includes testing standards, independent assessments, cybersecurity protections and requirements to report serious incidents. OpenAI also calls for written notice when a model circumvents security controls during evaluation. The Next Web describes the same package as an attempt to make existing practices at well-resourced frontier laboratories a national baseline. The request is directed at Congress; it is a company policy position, not legislation already introduced or enacted.[1], [2]
State bills remain the bridge until Congress acts
OpenAI says it will continue supporting state action while a federal framework is absent. The company cited four California bills covering independent safety assessments, auditor standards, protections for young people and safeguards against AI-enabled biological threats. Two measures on outside evaluations and auditors were signed on September 9. This state-level support does not replace the national rules OpenAI is seeking; the company presents it as an interim path while Congress considers a common requirement for advanced systems.[1], [2]
