Photo: Tima Miroshnichenko / Pexels
The White House AI advisor argues that OpenAI and Anthropic are using safety rhetoric to lock in their market dominance
David Sacks, the White House’s top AI advisor, is calling out the biggest names in artificial intelligence for what he sees as a cynical play dressed up in safety clothing. On September 13, Sacks publicly criticized OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei for backing a movement called “Pace The Frontier,” a coordinated push to slow the development of advanced AI systems.
What ‘Pace The Frontier’ actually proposes #
The initiative launched on July 28 as an open letter signed by over 1,134 individuals from major AI organizations, including employees at Google DeepMind, Meta, OpenAI, and Anthropic. By mid-September, that number had swelled past 1,300 signatories.
The letter didn’t demand an immediate halt to AI research. Instead, it called on governments to support tools enabling coordinated international policymaking aimed at slowing the pace of frontier AI development.
The movement emerged from growing safety concerns in the AI industry, particularly after incidents that raised alarms about the autonomy and unpredictability of advanced AI systems. Product liability risks, including those highlighted by episodes involving platforms like Hugging Face, added urgency to the conversation.
The bipartisan FRONTIER Act, introduced in July, reflects the broader political appetite for some form of AI regulation.
Sacks’ regulatory capture argument #
Sacks framed the regulatory proposals tied to “Pace The Frontier,” including requests for antitrust exemptions that would allow competitors to coordinate on safety standards, as textbook regulatory capture.
His logic runs like this: OpenAI and Anthropic already hold significant market share in frontier AI, with rapid revenue growth cementing their positions. If governments create international frameworks that restrict the pace of development, the companies already at the frontier get to stay there while everyone else is told to slow down before they arrive.
Sacks also noted that China is unlikely to participate in any coordinated global effort to slow AI development, making the whole exercise potentially self-defeating from a national competitiveness standpoint.
Rather than government-mandated pacing, Sacks argued that developers should voluntarily prioritize reliability and predictability in their products. He pointed to market pressures already pushing in that direction, suggesting that customers are increasingly favoring AI systems that work consistently over those that push capability boundaries at the expense of dependability.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our