By Paul Steidler, for Real Clear Markets

One of the most important and compelling aspects of America’s story is that for 250 years people have been free to innovate and produce new tools and technologies that make essential goods and services available at much lower cost, spreading prosperity to all. This consistent track record has meant unprecedented wealth and opportunity, the elimination of once horrific and commonplace fatal diseases, and dramatically improved affordability.

Today, artificial intelligence (AI), a technology that is more profound in its capacity to change and improve lives than electricity, holds the promise of transforming America. Indeed, it can take us to the golden age President Trump envisions.

Food, as it has for generations, will be produced more abundantly and require a smaller percentage of household budgets. Medical innovations will improve diagnoses and reduce illnesses. Even government regulation can be streamlined as AI ferrets out duplicative and outdated bureaucratic mandates.

One of the biggest challenges to realizing these benefits, however, is a spate of state laws that would raise AI development costs and perhaps even constrain new AI products and services, especially among mid-size and entrepreneurial companies.

With President Trump expected to issue an Executive Order this week to preempt state laws and restrictions on AI, there has been considerable discussion of the power dynamics between the Administration and blue states. Without an executive order and/or new laws, California and New York will exert major influence on national AI policy. A 50-state patchwork also raises costs in AI development and delays the deployment of new technologies.

While these measures have been blasted by the White House, and appropriately so, for raising AI development costs, creating marketplace uncertainty, and promoting DEI, there is a more fundamental issue: consumers will have higher costs due to delayed innovations. AI state regulation harms affordability.

More than 1,000 state AI bills have been introduced in 2025. According to the National Association of State Legislatures, as of July, “Thirty-eight states adopted or enacted around 100 measures this year.” Hundreds more remain under consideration.

These new laws and proposals include reporting requirements, mandates to use clean energy, AI model design specifications and restrictions, and, most notably, bias mitigation, i.e., DEI-related requirements or ensuring the absence thereof. Both red and blue states have numerous seemingly conflicting proposals in this regard.

There is some question as to whether many or all of these measures will withstand judicial scrutiny under the Commerce Clause of the U.S. Constitution, which gives Congress broad power to regulate interstate commerce and restricts states from impairing interstate commerce. The Internet and accompanying AI clearly constitute interstate commerce, given their use and dispersion throughout the U.S. and even worldwide.

Furthermore, AI and its uses are also already significantly governed. For example, you cannot use AI to steal money from a financial institution any more than you could rob a bank with guns. And if you use AI to scam an individual or harm his or her reputation, the consequences are severe.

The Administration’s approach of encouraging innovation and monitoring issues that arise is the best way to cultivate AI development and its benefits. If America had dozens of states crafting legislation on how cotton gins, tractors, steelmaking, electricity, autos, and more, we would have unduly delayed, if not killed, those wonderful developments.

America needs to deploy AI and have a state AI moratorium. We cannot afford not to do so.

Paul Steidler is a Senior Fellow with the Lexington Institute, a public policy think tank based in Arlington, Virginia.