A White House Senior Official said we should "Regulate AI Like We Do Pharmaceuticals," but the next day the White House said he was taken out of context

According to BotSeer Beating monitoring, on May 4th, The New York Times reported that the White House was discussing establishing a pre-release review mechanism for AI models, which was then in the "consideration" stage. Two days later, the situation escalated: White House National Economic Council Director Kevin Hassett directly appeared on Fox Business on Wednesday, stating that the government is exploring an executive order requiring government review before AI model deployment, "similar to FDA approval of drugs." Trump's first move in office was to revoke Biden's AI security executive order, and now he is pushing for a stricter pre-screening process.

Within less than 24 hours, the White House urgently toned down the situation. A senior official stated on Thursday that Hassett's words were "taken out of context," clarifying that the White House is seeking a "partnership" with businesses rather than "government regulation." White House Chief of Staff Susie Wiles made a late-night post on X on Wednesday to calm the situation, stating that the government "does not pick winners and losers," and that innovation should be driven by innovators rather than bureaucrats. This was her fourth post since creating the official account.

Although the tone softened, actions continued. Three sources familiar with the matter informed Politico that the White House is discussing having intelligence agencies conduct pre-release assessments of models before public deployment, with one aim being to ensure that the U.S. intelligence community can study and utilize these tools before Russia and China are aware of new capabilities. Deputy Secretary of Defense Emil Michael openly supported pre-assessment at an AI conference in Washington on Thursday, stating that Mythos is fundamentally a cybersecurity issue and that "these models will come sooner or later," requiring the government to establish response mechanisms. The Department of Commerce's AI Standards and Innovation Center (CAISI) had already signed a voluntary evaluation agreement for 2024 with OpenAI and Anthropic, expanding this week to include Google DeepMind, Microsoft, and xAI.

The immediate trigger for this policy reversal was Anthropic's disclosure last month of Mythos's vulnerability research capabilities being too strong to be publicly released. However, the White House found itself in a dilemma: on one hand, Trump signed an executive order prohibiting federal agencies from using Anthropic products and called its executives "left-wing crazies"; on the other hand, various federal agencies were eager to access Mythos to detect vulnerabilities in their systems.

The industry's reaction to mandatory pre-screening was significant. Information Technology and Innovation Foundation President Daniel Castro said, "If approval can be delayed arbitrarily, that's a problem for any company. The market access delay of weeks or months due to one company receiving approval while another does not has a significant impact."