AMERICA’S SPY AGENCIES WANT TO CONTROL AI — AND THEY ARE FIGHTING THE WHITE HOUSE FOR THE POWER TO DO IT A battle is underway inside the federal government over who gets to regulate artificial intelligence, and the stakes are not abstract. The Washington Post reported this week that the Trump administration is sharply divided over a proposal that would hand U.S. intelligence agencies a far larger role in evaluating advanced AI models before they reach the public. The intelligence community wants in. Other parts of the administration are pushing back hard. The push is being driven by what the government has seen from Anthropic’s Mythos model and other frontier AI systems now capable of identifying and exploiting software vulnerabilities at scale. Officials who have been briefed on these capabilities want a formal mechanism for evaluating what AI can do before it leaves the lab. The intelligence community argues it is best positioned to make those assessments and should be given the authority to act on them. Commerce Department officials and White House technology policy advisors are resisting. Their argument is that handing intelligence agencies oversight of commercial AI would build a regulatory framework that slows development and places American companies at a structural disadvantage against China at the worst possible time. The structure of this argument is familiar but the intensity is new. For two years the Trump administration dismissed safety regulation as a burden inherited from Biden. Now the government has watched AI find zero-day vulnerabilities in live systems and the conversation has changed entirely. Who wins this internal fight will determine whether the next generation of frontier AI gets a government minder before it ships. Keywords: AI regulation intelligence agencies, Trump AI oversight battle, US AI governance 2026, AI national security policy