OpenAI Seeks Congressional Shield for AI Safety Slowdown Amid Antitrust Fears

OpenAI has approached members of Congress to determine if coordinating an industry-wide slowdown of frontier AI models would violate federal antitrust laws. The move highlights a growing tension between the urgent push for AI safety and strict regulations designed to prevent market collusion.

According to TechRepublic, OpenAI sought specific guidance on whether mutual agreements to throttle the rollout of new capabilities could be interpreted as illegal output restrictions under the Sherman Antitrust Act. While leaders from OpenAI, Anthropic, and xAI have expressed a shared need to pace the development of frontier models to ensure safety, they fear that formalizing this coordination could trigger severe legal penalties. Sam Altman, CEO of OpenAI, has publicly aligned with this view, stating, "I agree with Dario [Amodei] that we need to pace the frontier."

The Legal Conflict

The current dilemma stems from the Sherman Antitrust Act, which prohibits competitors from agreeing to limit production or restrict output to manipulate a market. In the context of AI, a collective agreement to delay the release of a more powerful model for safety reasons could be viewed by regulators as an illegal attempt to stifle competition or maintain a dominant market position.

Some critics argue that these legal concerns are overstated. John Schulman noted that while antitrust law prohibits certain agreements, it does not prevent companies from jointly developing a proposal for safety standards. Similarly, Rep. Josh Gottheimer has pushed back against the need for a collective shield, suggesting that if leaders are truly concerned about safety, they can "pump the brakes at their own labs, today."

Proposed Legislative Solution

To resolve this impasse, a bipartisan bill known as the Collaboration on Adversarial Threats and Security Risks Act has been proposed. Sponsored by Senators Adam Schiff and Jim Banks, the legislation would provide a legal safe harbor for AI companies to collaborate on safety measures, provided they give advance notice to the Department of Justice. The bill is currently pending in committee.

Industry Implications

The outcome of this legislative effort carries significant weight for the future of the AI ecosystem. If a government-approved coordination framework is established, it could enable the industry to implement critical safety guardrails without the threat of litigation. However, there is a secondary risk: such a framework could inadvertently create a "dominant cartel." Critics worry that established incumbents might use safety coordination as a pretext to block newer, smaller competitors from advancing their capabilities, effectively protecting their own market share under the guise of security.

What's Next

Attention now turns to the congressional committee reviewing the Collaboration on Adversarial Threats and Security Risks Act. Whether the bill passes—and how narrowly it defines "safety collaborations"—will determine if the AI industry can move toward a synchronized pace of development or if the fear of antitrust prosecution will keep safety efforts fragmented.