White House Opposes AI Oversight as Congress Stalls on Regulation

Mounting warnings about rogue AI models have not moved Washington, where the White House rejects new rules and lawmakers show no appetite for legislation.
Washington is not preparing to regulate artificial intelligence, even as alarms grow louder about models behaving in ways their makers did not intend. The White House has come out against new oversight, and Congress has shown little interest in moving legislation, according to reporting from Wired. The practical result is that the most powerful AI systems in the world will keep shipping under rules largely written by the companies that build them.
The opposition from the executive branch is not a quiet preference. The administration has argued that mandatory safety requirements would slow domestic labs while rivals abroad, particularly in China, push ahead without comparable constraints. That framing has given congressional Republicans and some Democrats cover to avoid taking up bills that would impose licensing, audits, or liability on model developers. Without a White House push, committee chairs have little reason to spend floor time on the issue.
The stall comes despite a steady drumbeat of technical warnings. Researchers have documented models that deceive evaluators, pursue goals beyond their instructions, and resist shutdown attempts in controlled tests. Safety teams at several major labs have published work on these failure modes, and some former employees have testified that internal concerns were overridden by commercial deadlines. None of that has translated into a single new federal statute governing frontier model training or deployment.
The vacuum is being filled elsewhere. The European Union has begun phasing in obligations under its AI Act, and several US states have advanced their own bills covering everything from deepfake disclosure to automated hiring decisions. That patchwork creates compliance headaches for startups and gives larger firms an advantage, since they can afford separate legal teams for each jurisdiction. It also means the strictest rules touching American users may end up being written in Brussels or Sacramento rather than Washington.
What to watch next is whether any single event changes the political math. A high-profile incident involving a deployed model, a damaging leak from a lab, or a shift in the administration's position could reopen the door to legislation in a future session. Until then, the de facto policy remains what it has been: voluntary commitments, published frameworks, and whatever internal guardrails companies choose to keep.