Who Gets to Define the Rules for AI?

Aidan Gomez, the co-founder and CEO of Cohere, argues that a select group of dominant AI companies in Silicon Valley should not have the power to define the rules and safety standards for this groundbreaking technology. He asserts that similar attempts in the past, such as the establishment of Nationally Recognized Statistical Rating Organizations and the Motor Vehicle Block Exemption, have resulted in market structures that protect incumbents and limit competition, all under the guise of public safety.

Gomez emphasizes the importance of guardrails for AI but emphasizes that the real debate should be about who gets to write them and who benefits from the rules. He draws parallels to previous monopolies in the bond rating and car manufacturing industries, where incumbents set the standards to their own advantage, often without input from other stakeholders.

Gomez argues that while independent review of AI systems is a good idea, the Anthropic CEO's roadmap raises fundamental questions about who writes and oversees the standards, who gets to participate in the decision-making process, and whether a handful of powerful labs should be granted a narrow antitrust waiver to coordinate with each other.

He contends that adding more participants does not solve the issue, as the real problem is the lack of input from other companies, governments, and citizens in the creation of AI regulations.