OpenAI is exploring whether a coordinated slowdown of AI development among major labs would be legal under current antitrust laws, as the urgency to safeguard rapidly advancing AI technology grows.
- Coordinated AI slowdowns risk violating US antitrust laws
- Bipartisan bill seeks to permit safety collaboration legally
- Industry rivalry and national security concerns complicate consensus
What happened
OpenAI's chief scientist recently proposed that the AI research community should coordinate to slow the pace of AI development to ensure safety, anticipating that voluntary slowdowns will become common until safety benchmarks are agreed upon. This suggestion comes amid growing alarm over rapid advancements in AI capabilities and concerns about security incidents linked to powerful models.
However, legal experts warn such coordination between competing AI labs could potentially violate the Sherman Antitrust Act, which prohibits agreements that restrict output. Despite these concerns, some lawmakers introduced legislation aimed at allowing AI companies to collaborate on security and safety issues without fear of antitrust repercussions, although the bill has yet to advance in Congress.
Why it matters
The prospect of an industry-wide slowdown coordinated for safety reasons highlights a critical tension between rapid innovation and regulatory control in AI development. Antitrust law that traditionally prevents collusion among companies could inadvertently hinder necessary cooperation to mitigate risks posed by advanced AI systems.
At the same time, competing interests persist within the AI industry: companies are racing to dominate future markets and some view leadership in AI as a national security imperative, particularly in relation to China. Differing perspectives on what constitutes safe AI development also create barriers to collaboration, making legal clarity essential to balance safety goals with competitive dynamics.
What to watch next
The progress of the bipartisan Collaboration on Adversarial Threats and Security Risks Act will be key to enabling safer AI innovation. Its passage could provide the legal framework for AI labs to jointly address safety without antitrust risk, although timing is uncertain amid the US midterm elections and various political priorities.
Industry dynamics will also be important to monitor—whether major AI developers like OpenAI and Anthropic can overcome internal rivalries to jointly propose and implement pacing measures. Additionally, ongoing public risks and incidents involving AI systems may increase pressure for regulatory action and clearer guidelines for collaboration on AI safety.