Aidan Gomez, Co-Founder & CEO of Cohere, which builds enterprise artificial intelligence tools and large language models designed specifically for businesses and regulated industries, argues that AI needs strong safety rules, but he believes a small group of dominant Silicon Valley companies should not control them. He supports independent testing and oversight, while warning that giving major AI companies the power to create industry-wide standards could reduce competition and allow those companies to protect their existing market positions.
Gomez proposes an evidence-based and internationally developed safety framework. He calls for greater transparency from AI developers, independent testing based on specific risks, and assurance mechanisms similar to those used in industries such as finance, aviation, and nuclear energy. In his view, regulations should focus on what an AI system can actually do and where it is deployed, rather than simply on the size or resources of the company that created it.
He also argues that policymakers should focus on realistic and immediate risks, such as cyberattacks, fraud, voice cloning, critical-infrastructure failures, and misuse of AI in sensitive environments, rather than relying primarily on speculative predictions about superintelligent AI destroying humanity. His central message is that AI safety should involve governments, researchers, smaller companies, civil society, and other stakeholders so that the rules are developed through an open, evidence-based process rather than being determined by a small group of companies.
