OpenAI is urging the U.S. Congress to pass mandatory, national AI safety requirements before it adjourns, according to multiple reports. The company says it supports capability-based regulation and is calling for common testing and safety processes for advanced AI systems.
Several outlets describe the proposal as outlining a set of requirements OpenAI says it wants to meet itself and wants applied to other companies. Coverage also links the push to broader concerns about how AI systems behave during testing, including risks when models access external systems and potentially produce unexpected results.
While the outlets agree on the thrust of the request—legislation in the United States and a standardized approach to AI safety—details of emphasis vary. Some focus on the regulatory approach and OpenAI’s timing with Congress, while others highlight the specific kinds of testing and external access concerns that have prompted calls for tighter oversight.
