David Sacks Accuses AI Giants of Using Safety Pleas for Regulatory Capture

David Sacks, Chair of the President's Council of Advisors on Science and Technology, has challenged calls for government-mandated AI regulation from industry leaders. Sacks argues that if frontier labs believe their models pose significant risks, they should exercise voluntary restraint rather than lobbying for state intervention.

Responding to pleas from OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei to "pace the frontier" of AI development, Sacks asserted that these companies do not require government regulations to slow their own progress if safety is the primary concern. He characterized the push for formal regulatory frameworks as a potential strategy for "regulatory capture," suggesting that such rules would protect established incumbents by creating insurmountable barriers for smaller competitors. Sacks specifically accused Anthropic of employing a "sophisticated regulatory capture strategy based on fear-mongering" intended to damage the broader startup ecosystem.

The Debate Over AI Pacing

This conflict emerges as frontier AI labs increasingly call for international authorities or state-level safety laws to manage the rate of growth in AI capabilities. While these labs frame their requests as essential safety measures, critics view the move as an attempt to "moat" their businesses. By advocating for compliance burdens that only well-resourced companies can afford, incumbents may effectively stifle open-source competition and new entrants.

Sacks further suggested that the motivation to slow development may not be rooted in pure altruism. He pointed to potential product-liability exposure, particularly regarding the role of AI in cyberattacks, as a more likely driver for the industry's sudden interest in pacing.

Industry Implications

The divide highlights a fundamental disagreement over the future of AI governance: whether safety is best achieved through voluntary corporate restraint, open-source competition, or government-mandated frameworks. Sacks has been blunt about the optics of these requests, stating that demanding a preferred regulatory framework as the price for safety "will look like blackmail of the public and the political system," adding, "So just do it."

If the industry shifts toward a government-certified model, the AI landscape could transform from an open ecosystem into a consolidated oligarchy of a few approved labs. This would fundamentally alter how innovation occurs, shifting power from agile startups to a small group of regulated entities.

What to Watch

As the debate intensifies, the focus remains on whether the U.S. government will adopt the regulatory frameworks suggested by frontier labs or lean toward the voluntary, open-competition approach advocated by Sacks. Observers are watching for further evidence of how proposed safety audits and enforcement mechanisms might impact the ability of smaller AI startups to compete in the frontier model space.