Aidan Gomez, the co-founder and CEO of Cohere, has released a perspective addressing the ongoing debate over AI governance. Gomez argues against a model where a small group of market-dominant AI companies from Silicon Valley defines the safety standards and rules for the entire world.

He specifically criticizes recent proposals, such as the roadmap published by Anthropic CEO Dario Amodei, which seeks government antitrust exemptions in the name of safety. Gomez suggests that these efforts risk creating a "cartel" that protects incumbents by setting high entry barriers—such as vast computing power and massive infrastructure requirements—that smaller developers cannot meet.

Gomez contends that true AI safety requires inclusive, evidence-based processes involving diverse stakeholders, including academics, civil society, and open-source developers. He emphasizes that safety rules should be designed to reduce risk across the entire field rather than serving to preserve the commercial advantages of existing industry leaders.

Sources: