Over 270 Organizations Sign Open Weights AI Letter

More than 270 companies and organizations signed the Open Weights and American AI Leadership letter, arguing that open-weight models increase access, competition, and organizational control over AI deployment.

Anthropic declined to sign the letter, maintaining that sufficiently capable models pose risks because safeguards can be removed once weights are released. The company is instead advocating for mandatory safety testing for both open and closed models.

In response to the debate, industry commentator Harini Gopalakrishnan proposed a system of regulated diffusion. This framework would allow frontier AI companies to monetize new models while providing a pathway for previous generations to become open-weight after passing independent safety assessments. Gopalakrishnan argued that safety certification should be handled by a consortium of researchers, regulators, and domain experts rather than developers, who may have conflicting economic incentives.