U.S. Presidential Science and Technology Advisory Committee Chair Criticizes OpenAI and Anthropic

On September 14, 2026, the chair of the U.S. Presidential Science and Technology Advisory Committee, David Oliver Sacks, publicly criticized OpenAI and Anthropic regarding their recent calls to slow down the development of advanced artificial intelligence. Sacks expressed his views on social media, pointing out that both companies dominate the advanced AI sector in terms of market share, revenue growth, and the capabilities of their AI models, effectively creating a duopoly in the field.

Sacks noted that while leaders from both companies, including Altman, have suggested a need to decelerate AI developments, he finds their motivations suspect. He questioned the legitimacy of their claims for slowing down, suggesting that if their unreleased models are indeed powerful enough to warrant caution, they should responsibly make that decision without seeking external validation. He warned against any attempts to pause antitrust laws to form a cartel or to rely on regulatory approvals that could undermine product accountability.

He further stated that the motivations behind their calls for a slowdown should not be framed as purely altruistic, especially if their products could lead to destructive cyberattacks, thus exposing them to significant liability risks. He remarked that the market is already punishing companies whose models behave unpredictably or without authorization.

Sacks referenced the Hugging Face incident, suggesting that both companies are trading some of their original capabilities for reliability or predictability, which he deemed a sound business decision. He indicated that this trade-off could be viewed as a form of alignment but fundamentally serves to meet customer demands.

The chair emphasized that a slowdown in advanced AI development could provide time for a more rational discussion on regulatory issues, beyond the extreme views of a complete halt as suggested by Senator Bernie Sanders. He also noted that the challenges posed by other countries, particularly in forming global agreements, must be considered.

Sacks concluded by asserting that if OpenAI and Anthropic wish to define their own regulatory framework as a condition for slowing their development, this would be tantamount to coercion within the public and political spheres. He urged them to act decisively, warning that failure to do so would reveal their actions as yet another attempt at regulatory intimidation or part of an electoral strategy.

Reported by ITHome.