Compromise Necessary but Not Sufficient for AI Preemption

This commentary argues that preempting state AI laws will need bipartisan compromise, even with all three federal branches engaged. It notes that Representatives Jay Obernolte and Lori Trahan released a draft Great American AI Act the previous week, which would set up a national regulatory framework and is the latest try at federal preemption.

The article traces the background. In November 2025 President Trump demanded one national standard rather than a patchwork of state regimes, and a month later signed an executive order on a national policy framework. The order set a March 11 deadline for agencies to act, yet only the Department of Justice complied, creating an AI Litigation Task Force on January 9. Commerce has not published its evaluation of state laws, which leaves the FCC unable to proceed, and the FTC has issued no policy statement. In April the task force secured its first success when, after xAI sued Colorado and the DOJ intervened, a court blocked enforcement of the Colorado AI Act. The author says litigation alone falls short because it addresses one law at a time.

The piece then describes the shift to Congress. An earlier attempt by Senator Cruz to add a moratorium to a budget bill lost 99 to 1, months before the order. Senator Blackburn's draft TRUMP AMERICA AI Act would override conflicting state laws, while the Future of AI Innovation Act from Senators Young and Cantwell would make CAISI central to federal authority, implying preemption. The Obernolte-Trahan bill blends both: it would preempt state rules aimed at AI model development for three years, but leave general laws and post-deployment rules untouched. It already faces resistance from a House commission and Anthropic. The author concludes that Congress is gridlocked and a patchwork seems inevitable.