White House AI czar Michael Kratsios unveils first comprehensive national AI legislative framework

Summary

Michael Kratsios, the White House AI czar, unveiled what he describes as the first comprehensive national AI legislative framework presented to Congress by a White House. The central problem it is trying to solve is regulatory fragmentation: dozens of states have passed or proposed their own AI rules, and while large companies can navigate that patchwork, small startups largely cannot.

National preemption

The framework's core ask is a single federal standard that preempts state-level AI legislation. Kratsios is explicit that laws like a proposed New York rule barring AI from giving legal advice are exactly what the framework is designed to stop. Federal preemption would prevent individual states from banning AI in specific verticals, whether legal services, healthcare, or anything else.

Ratepayer protection

The framework urges Congress to codify the ratepayer protection pledge that the Trump administration extracted from major AI companies, committing them to build, bring, or buy their own power for any data center they construct. The sharper ask is to enshrine in law the right to build behind-the-meter power generation, something the Biden administration had moved to restrict via executive action. Kratsios argues legalizing behind-the-meter generation unlocks a materially different range of options for powering data centers.

Child safety

Kratsios identifies child safety provisions as the most likely vehicle for bipartisan progress, describing it as an "80-20 issue." The framework asks Congress to require platforms to give parents visibility into what their children are seeing, who is interacting with them, and tools to control future exposure. His read is that the industry is already moving in this direction — Instagram parental controls, YouTube Kids — but that a national law is needed to make the standard permanent.

AI risk and national security

On AGI and safety evaluations, the administration is deliberately not building a regulatory body with the power to approve or ban models. Instead, the framework tasks national security agencies with developing internal expertise to evaluate models for national security purposes, supporting warfighters and cybersecurity rather than acting as commercial gatekeepers. Kratsios frames this as a clean break from the Biden-era approach, where safety testing was tied to potential bans or nationalization.

Workforce and small business

The administration's workforce argument centers on redirecting existing Labor Department, Education Department, and Small Business Administration programs toward AI readiness rather than creating new ones. Kratsios points to small businesses, which employ the vast majority of Americans, as the most underappreciated beneficiary of AI productivity gains — the one-to-ten person shop that can now handle inbound calls, client qualification, and scheduling through AI without hiring additional staff.

Kratsios says the administration's target is to get legislation passed within calendar year 2026, with child safety provisions as the most viable entry point for bipartisan momentum.