Cohere CEO Challenges Silicon Valley's Grip on AI Safety Standards

Aidan Gomez, the co-founder and CEO of Cohere, is questioning whether a small group of market-dominant Silicon Valley firms should define global safety standards for artificial intelligence. In a perspective piece, Gomez argues that the governance of a generational technology should not be left to a handful of select companies.

Writing in a post titled "Who Gets to Define the Rules for AI?", Gomez advocates for a shift away from closed-door governance. He argues that the process for establishing AI rules must be open and evidence-based, allowing for broader participation rather than being restricted to a few industry incumbents. Gomez explicitly asks if a small number of dominant firms from one region should define the rules and safety standards for the entire world.

The Debate Over Pacing

This push for open governance arrives amid a deepening divide within the AI industry regarding "frontier risks." The tension is highlighted by contrasting views from industry leaders. Dario Amodei, CEO of Anthropic, recently published an essay titled "We Must Pace the Frontier," in which he calls for a deliberate slowdown in the rate of AI capability gains to better manage potential catastrophic risks.

While some leaders view this "pacing" as a necessary safety precaution, others in the broader tech community suggest such moves could lead toward regulatory capture. In this scenario, the largest players in the market could influence legislation to create high barriers to entry, effectively preventing smaller competitors from entering the field under the guise of safety.

Why Governance Models Matter

The resolution of this debate will determine who controls the guardrails of one of the most transformative technologies of the century. If a few dominant companies successfully define the rules, they gain significant influence over the speed of innovation and the ethical boundaries of AI development globally.

Conversely, a democratic or open-source approach to governance could accelerate the adoption of AI across different sectors and geographies. However, such a decentralized model may face challenges in implementing unified safety standards quickly enough to mitigate systemic risks.

The Path Forward

As the industry continues to scale, the conflict between closed-door and open governance remains unresolved. Observers are now watching to see if international regulatory bodies will adopt the evidence-based, inclusive approach suggested by Gomez or if they will rely on the frameworks proposed by the current market leaders. Whether the global community can balance the need for rapid innovation with the necessity of rigorous, transparent safety standards remains the central question for AI policy.