OpenAI Pushes for Mandatory National AI Safety Rules in the U.S.

OpenAI is urging the U.S. government to establish mandatory national safety requirements for advanced artificial intelligence systems, saying voluntary commitments may not be enough as AI capabilities continue to advance.

The company is calling for a federal framework that would set clear requirements for testing, independent assessments, cybersecurity protections and reporting of significant AI incidents.

According to Reuters, OpenAI Chief Global Affairs Officer Chris Lehane said the United States needs mandatory, capability-based national regulation that can evolve as AI technology changes.

OpenAI Calls for Federal AI Safety Standards

The push comes as the United States continues to debate how advanced AI systems should be regulated.

OpenAI wants national requirements to focus on the capabilities and potential risks of AI systems rather than applying identical rules to every type of artificial intelligence.

Under the proposed approach, more advanced systems could face stronger safety requirements before they are deployed more widely.

The company says the framework should include standardized testing, independent evaluations and cybersecurity safeguards.

What Could the Proposed Rules Cover?

OpenAI's proposal includes several areas that could become part of future federal AI regulation.

These include:

• Safety testing for advanced AI systems
• Independent assessments of highly capable models
• Stronger cybersecurity protections
• Reporting requirements for significant AI incidents
• Capability-based safety standards
• Rules that can be updated as AI technology develops

The goal would be to create clearer expectations for companies developing increasingly capable AI systems while giving policymakers a more consistent national framework.

Why AI Safety Rules Are Becoming a Bigger Issue

AI systems are becoming increasingly capable and are being used across software development, business, research, education and other areas.

Some newer AI systems can also perform multi-step tasks with less direct human supervision. That has increased attention on how companies test, monitor and control advanced models.

Recent incidents involving AI systems accessing external websites and systems during testing have added to concerns about whether existing voluntary safety practices are sufficient. Reuters has reported several cases involving AI agents behaving in unintended ways during testing.

OpenAI has also acknowledged that the industry needs greater transparency around incidents involving unintended AI behavior and said there is not yet a clear standard for reporting such incidents.

Federal Rules Could Create a More Consistent Approach

One challenge facing the U.S. technology industry is the growing number of different AI-related rules being considered or adopted at the state level.

OpenAI has supported several California AI safety measures while continuing to call for a nationwide framework.

The company said it would continue supporting state-level legislation while encouraging Congress to establish federal requirements.

A national framework could potentially give developers clearer rules across the United States instead of requiring companies to navigate different standards in different states.

OpenAI Also Supports Independent Evaluations

Another important part of the proposal is independent assessment.

Independent evaluations could provide an outside review of advanced AI systems instead of relying entirely on companies to evaluate their own products.

Reuters reported that OpenAI has also endorsed California measures involving independent third-party evaluation and audits of AI systems.

For policymakers, independent testing could become an important part of determining whether highly capable AI systems meet safety expectations before broader deployment.

What This Means for the U.S. AI Industry

The debate could have significant consequences for American technology companies.

Clear national standards could make compliance more predictable and potentially reduce uncertainty caused by different state-level requirements.

At the same time, companies and policymakers will have to consider how regulations could affect innovation, development costs and the ability of U.S. companies to compete internationally.

The challenge for Congress will be finding a balance between encouraging AI development and creating safeguards that keep pace with technological progress.

What Happens Next?

OpenAI is urging Congress to establish national AI safety requirements and act before the end of the current congressional session.

Congress is scheduled to adjourn in December, according to Reuters. However, there is no guarantee that lawmakers will reach agreement on a comprehensive federal AI safety framework before then.

The debate is likely to continue as AI companies develop more capable systems and lawmakers consider how existing laws and new regulations should apply to them.

For the United States, the central question will be how to encourage AI innovation while ensuring that advanced systems are developed and deployed with appropriate safety, security and accountability measures.

Frequently Asked Questions

What is OpenAI asking Congress to do?

OpenAI is asking Congress to establish mandatory national safety requirements for advanced AI systems, including testing, independent assessments, cybersecurity protections and incident reporting.

Why does OpenAI want national AI safety rules?

OpenAI says voluntary commitments may not be sufficient as AI systems become more capable. The company wants a consistent federal framework that can evolve with the technology.

Would the proposed rules apply to every AI system?

OpenAI is calling for capability-based requirements. This approach would generally place stronger requirements on more advanced and capable AI systems.

What role could independent AI assessments play?

Independent assessments could provide outside evaluations of advanced AI systems and help determine whether they meet established safety standards.

When could the U.S. introduce new AI safety requirements?

There is currently no guaranteed date for comprehensive federal legislation. OpenAI is urging Congress to act before its December adjournment, but any legislation would still need to go through the congressional process.

Related U.S. AI Coverage

Readers interested in the broader U.S. debate over AI safety can also read our coverage of U.S.-China AI safety talks and cybersecurity risks.

Sources

Reuters: OpenAI pushes for mandatory national AI safety rules.

Reuters: AI models' capabilities leap comes with new safety warnings.

Disclaimer

This article is for informational purposes and summarizes publicly reported developments. AI policy proposals and regulatory discussions can change as lawmakers, regulators and technology companies continue their work.