Quick Take

Trump’s Executive Order on AI Creates a Dangerous Regulatory Vacuum

Luke Barnes

December 16, 2025

Last week, tech companies saw an early return on their efforts to cultivate ties with Donald Trump.

On December 13, the President signed an Executive Order severely limiting the ability of states to regulate Artificial Intelligence (AI). The order, which states that “[AI] companies must be free to innovate without cumbersome regulation,” allows the attorney general to sue and/or withhold funds from states that act outside of a “minimally burdensome national policy framework.”

The Order, however, is already facing pushback. California Governor Gavin Newsom described the policy as a “con” in the immediate aftermath of the signing, and the state’s Attorney General promised to examine its legality. Legal scholars were also wary about its legal standing. As Fordham University Law Professor Olivier Sylvain noted, federal agencies cannot challenge state regulations that have been lawfully enacted unless explicitly mandated to do so by Congress.

But the question is less about whether the specific Executive Order would survive the courts, and more about the signal it sends to the AI industry. In effect, it seeks to disable any meaningful guardrails on AI development—however imperfect they might be— without offering a viable replacement.

In the absence of comprehensive federal legislation, states have been at the forefront of regulating the fast-moving AI landscape. In September, for instance, California passed Senate Bill 53 (SB 53), which requires AI developers to publish updates on their security protocols and provides whistleblower protections for employees who flag risks. By late 2025, at least 38 states (including California) had enacted some form of AI-related legislation, reflecting the urgency of the issue and the federal legislative vacuum.

The Trump administration and its allies are not wrong to argue that a 50-state patchwork of regulation creates compliance challenges for companies operating at national or global scale, particularly in an industry moving at such speed. But that does not absolve them of their failure to offer a workable alternative for AI regulation.

The current Executive Order does not replace the existing state laws with any sort of coherent federal framework. Instead, it creates a legal limbo in which companies can argue that compliance with state-level safeguards is optional. This opens the door to well-worn corporate evasion tactics, including strategic delay, forum shopping, or regulatory chilling. This structure, in turn, favors the AI firms that have the resources to fund expansive legal teams—in other words, the same actors already dominating the AI marketplace—who have already drawn criticism for their projects, like ChatGPT’s deliberately sycophantic behavior.

But this short-sighted vision may have political consequences. As the Washington Post recently reported, some conservatives increasingly view Trump’s courtship of Silicon Valley investors as a betrayal of his populist base. “It feels like millions of votes across the country just got traded for thousands of [venture capitalist] and tech-rich votes in regions Republicans will never win,” one source told the paper.

This points to the potential of an unlikely convergence to counter Trump’s ban on state-level AI regulation. Progressives alarmed by unchecked corporate power and right-wing populists irate at Trump’s acquiescence to the tech-elite are identifying the same policy failure.