States Can Authorize AI Companies to Cooperate on Safety Without Violating Antitrust Laws
State governments possess legal authority under Parker immunity doctrine to permit AI companies to coordinate on safety measures without breaching federal antitrust laws. The principle, established in a 1943 Supreme Court case, allows states to regulate commerce within their borders and effectively supersede Sherman Act requirements. This pathway could enable AI firms to slow development for safety purposes without requiring federal antitrust waivers.
Parker immunity stems from a 1943 Supreme Court decision that permitted California to regulate raisin production and pricing, effectively overriding federal competition law. The doctrine recognizes states' sovereign authority to establish their own commercial rules within their borders. A later Supreme Court ruling refined this into a two-part legal standard: regulations must explicitly state the state's policy and include active government oversight of compliance.
The current debate centers on whether AI companies can coordinate safety practices without facing federal antitrust violations. Leading AI firms worry that slowing development collaboratively could expose them to Sherman Act liability, prompting requests for federal waivers. The Parker immunity pathway suggests state governments could instead authorize such cooperation through legislation, potentially bypassing the need for federal intervention.
If states adopt Parker immunity frameworks for AI safety coordination, it could reshape how the industry develops advanced systems. AI companies might gain legal cover to collaborate on safety standards and testing protocols, potentially reducing competitive pressure to release untested models rapidly. However, this approach could also concentrate regulatory authority at the state level, create compliance complexity across jurisdictions, and raise questions about whether state-level oversight adequately addresses risks from technology with national and global implications.