Gomez: Silicon Valley is trying to turn AI safety into a cartel

Aidan Gomez, co-founder and chief executive of Cohere, has launched a sharp attack on the effort by the largest AI labs to write the safety standards and rules that would govern the entire industry. He argues that a handful of dominant Silicon Valley companies are using public anxiety about generative technology to bend competition rules and dictate the pace and terms for everyone else — a move he describes as "a wolf in sheep's clothing, a cartel by any other name."

Gomez is not speaking from the sidelines. Cohere runs models inside banks, telecommunications networks and defence ministries — environments where a failure can disrupt financial systems, critical infrastructure and essential services at scale. He stresses that the capabilities that find vulnerabilities in your code are the same capabilities that find vulnerabilities in everyone else's, and that cyber offence is becoming cheaper faster than defence is improving. That gap, he says, should worry more than just his company.

To illustrate where a monopoly on standard-setting leads, Gomez points to 1975. The US Securities and Exchange Commission (SEC) needed reliable bond ratings for its capital rules and designated three firms as Nationally Recognized Statistical Rating Organizations (NRSRO) without publishing criteria for new entrants. The result: 25 years later only three agencies still existed, funded by the issuers they rated, and they handed out AAA ratings to subprime mortgage-backed securities — a decision that nearly brought down the global economy.

Europe repeated the mistake in 1985. Carmakers lobbied for a blanket antitrust exemption, the Motor Vehicle Block Exemption, arguing that modern vehicles were complex, safety-critical machines requiring control over who could sell and service them. The regulation let manufacturers set standards for facilities, equipment and training, ostensibly for safety. In practice no safer cars emerged, and it took roughly 25 years of reform to dismantle the monopoly and prove that strict safety standards can be enforced without handing the keys to vested interests.

Gomez makes clear that the need for guardrails is not in dispute. The fight is over who writes them, who gets a seat at the table, and whose interests the rules serve. The question, in his framing, is whether we allow free choice grounded in scientific evidence, or hand the reins of the most consequential technology in human history to a handful of Silicon Valley executives.