Pritzker urges Congress to set federal AI rules without overriding state protections

Illinois Gov. JB Pritzker is asking the incoming Democratic Congress to make AI safety, digital privacy and data-center efficiency federal responsibilities. His proposed agenda would add national requirements while preserving state protections—a position that puts the relationship between Washington and state governments at the center of his call for regulation.

The proposal follows Illinois’ own move into AI oversight. In July, the state joined New York and California in passing legislation governing frontier AI, the most advanced models. Illinois’ Artificial Intelligence Safety Measures Act includes mandatory annual third-party audits of frontier models, giving Pritzker an existing state framework to point to as he presses for federal action.

Three federal duties on Pritzker’s agenda

Pritzker included the digital protection framework in a suggested 200-point national agenda for the incoming Congress, Capitol News Illinois reported on November 9. The request brings together three areas of policy rather than focusing only on the risks posed by advanced AI models:

• Digital privacy: Establish durable federal privacy laws that protect Americans’ personal data while preserving economic development.
• AI safety: Require risk assessments, testing and monitoring at the federal level, making those activities requirements rather than leaving them solely to company discretion.
• Data centers: Require energy and water efficiency, putting the infrastructure supporting digital services alongside privacy and AI safety in the proposed framework.

Pritzker’s condition is that federal action should build on state leadership, not override measures Illinois and other states have already taken. His agenda is a policy request, not an enacted federal framework. Whether Congress will act, and whether President Donald Trump would sign resulting legislation, remains uncertain.

A national framework, but room for states

The appeal for federal rules has support from people with different concerns about state regulation. Illinois state Rep. Daniel Didech told Capitol News Illinois that robust federal rules addressing catastrophic AI risks would be preferable to a patchwork of 50 state laws. He argued that states have been filling gaps left by federal inaction.

Megan Stokes, state director at the Computer and Communications Industry Association, also favored a federal solution. Her concern is the cost of complying with conflicting state requirements. She said that problem reaches beyond frontier models to chatbots and applications in areas such as education and therapy.

When state laws start to conflict, and then you have this really like costly compliance landscape that’s going to hurt startups

Megan Stokes, Computer and Communications Industry Association, speaking to Capitol News Illinois

Audits expose the implementation dispute

Illinois’ audit requirement shows where agreement on oversight gives way to disagreement about execution. Didech argued that independent reviewers are needed because developers both write their safety plans and judge whether those plans are sufficient. He compared the government’s role with public-safety oversight of other powerful technologies.

Stokes raised questions about auditors’ training, independence, protection of industry secrets and treatment of whistleblowers. She also questioned what the auditing model would look like. Her concern was not simply the burden of a review: an audit could give the public a sense of assurance that its actual scrutiny does not justify.

There is also a question of which harms regulation should cover. DePaul University professor Bamshad Mobasher said protections should address undue surveillance, discrimination and leaks of private data. He warned that state legislation, including Illinois’ approach, concentrates on larger companies rather than necessarily addressing harms from applications built using their technology.

Timing presents another challenge for lawmakers. University of Illinois Urbana-Champaign professor Volodymyr Kindratenko cautioned that technology can evolve in unexpected directions, making today’s proposed guidelines less relevant tomorrow. That leaves the incoming Congress facing questions not only about whether to regulate, but how to write requirements that remain useful as AI changes.