Anthropic CEO Rejects Claim That AI Regulation Concentrates Power

Anthropic CEO Dario Amodei has pushed back against the notion that regulating artificial intelligence would inevitably consolidate power among a few large companies and governments. He argued that thoughtfully crafted rules could instead limit the influence of frontier AI firms while giving smaller competitors a better chance to catch up.

Amodei was responding to comments from investor Gavin Baker, who suggested that AI risks might be better addressed by distributing AI capabilities widely rather than through regulation that could concentrate control. Amodei called this a "false choice," insisting that fair institutional processes can check corporate power and protect individuals. He said regulation can be designed to place heavier obligations on the most advanced AI companies than on smaller challengers.

As an example, Amodei pointed to Anthropic's support for California's SB53 and its earlier stance on SB1047, both of which exempted companies below certain revenue or model-training thresholds. He also noted that Anthropic has backed testing frameworks that impose stricter requirements on frontier models than on less advanced systems, a move he believes could benefit challengers, including developers of open-weight models.

Amodei acknowledged that AI is structurally prone to concentrating power due to the economics and computing demands of scaling advanced models. While open-weight models can help spread capabilities, he said they are not enough because access to large-scale computing and chips remains concentrated among frontier firms and major hardware providers.

He also rejected the suggestion that his public messaging on AI has been overly negative. Amodei said he has tried to balance warnings about risks with the technology's potential benefits, citing his essay "Machines of Loving Grace," which explored AI's promise in healthcare and biology. He argued that the industry's bigger problem is a crisis of public trust, and that companies should earn confidence by delivering tangible benefits rather than relying on optimistic marketing.

Amodei remains supportive of pre-deployment testing for frontier models and testing of open-weight models as they approach the frontier. He also expressed support for a proposal by Google DeepMind CEO Demis Hassabis to create an AI oversight body modeled on FINRA, the financial industry regulator.