How Safety Collaboration Might Conflict With Existing Laws
OpenAI has asked Congress to clarify whether efforts by competing AI labs to coordinate on safety measures might violate antitrust laws. The company says it wants to work with rivals on responsible development but worries legal barriers could block such cooperation. This comes as AI systems grow more powerful and concerns about risks increase.
Breaking news
Tracking AI Search Visibility for Enterprise Brands
Singapore Expands Semiconductor Sector Amid AI Surge
Insurance Tech Startup Outmarket Secures $34.5 Million in Series B Funding
DeepSeek Partners with Huawei to Build Programming Tools for Ascend ChipsOpenAI argues that sharing safety practices, testing methods, and risk assessments across companies could help prevent harmful outcomes from advanced AI. However, current antitrust regulations were designed to stop price-fixing and market manipulation, not to govern technical collaboration on safety. The firm fears that well-intentioned coordination could be mistaken for anti-competitive behavior, discouraging labs from working together. It wants lawmakers to define clear boundaries so safety efforts do not trigger legal scrutiny.
What Specific Protections Do AI Labs Need From Congress?
Antitrust laws in the U. S. prohibit agreements between companies, even if aimed at public good, can raise red flags if they involve sharing sensitive operational details. OpenAI notes that discussing model evaluations, red-teaming results, or alignment strategies could be seen as exchanging competitively valuable information. While no enforcement actions have occurred yet, the uncertainty creates a chilling effect. Labs may hold back on sharing critical safety insights to avoid investigations. OpenAI believes this hesitation could slow progress on making AI systems more reliable and aligned with human values.
OpenAI is not seeking immunity from antitrust oversight but rather guidance on what types of safety coordination remain permissible. It suggests creating a safe harbor for activities like joint benchmarking, vulnerability disclosure, or sharing best practices for monitoring AI behavior. The company points to similar arrangements in aviation and nuclear industries, where competitors collaborate on safety without violating competition rules. Clear congressional guidance, OpenAI says, would allow labs to act responsibly without fear of unintended legal consequences.
Why is OpenAI concerned about antitrust laws affecting AI safety work? OpenAI fears that sharing safety-related information with rivals could be interpreted as anti-competitive collaboration under current laws, potentially discouraging cooperation even when intended to reduce risks.
Frequently Asked Questions
What kind of coordination does OpenAI want to protect? The company wants to safeguard efforts like joint safety testing, sharing risk assessment methods, and developing common evaluation tools, which it believes improve overall AI safety without harming competition.
Has any AI lab been penalized for safety collaboration so far? No enforcement actions have been taken against AI labs for safety-related coordination, but OpenAI says the lack of clarity creates uncertainty that may already be influencing behavior.
