Why Congress Should Pass a National AI Regulatory Framework

We are sliding into the waning days of the 119th Congress, with only around 30 days left in the legislative calendar for the House of Representatives and about 40 days left for the Senate before the election.

This therefore seems as good a time as any to recall why it is that having a national framework for AI regulation is so important.

Our Constitution grants Congress the unique right (and, we would argue, sometimes even the responsibility) to regulate interstate commerce.

Many topics are clearly not the legitimate domain of federal regulation; it is good and reasonable that we have fifty states competing and experimenting in areas ranging across the spectrum from sales tax rates to public school curricula. This flexibility is often spoken of, fondly, as the “laboratory of democracy”.

However, there are times when the rules and regulations of individual states do not stay comfortably within their own borders. We would not be as fond of a state legislature imposing de facto national rules on topics like monetary policy or national defense.

Indeed, Congress has chosen over the decades to establish national policies on topics as diverse as airline ticket pricing, radio spectrum regulation, telecommunications, employee benefit plans, and a host of other topics where it was (usually correctly) judged that national standards were required rather than a minefield of individual state laws.

The goal, in most cases, has been to provide a clear and understandable standard for people attempting to conduct business that is almost inherently interstate in nature.

Artificial intelligence systems are just in their infancy, roughly where the Internet was in the early to mid-1990s. However, the area has attracted a great deal of attention from the public, the press, and investors, and therefore naturally has also attracted the attention of state legislators. Thousands of local regulatory regimes have been proposed for the industry, many involving conflicting or outright contradictory requirements. Many more involve standards that are simply impossible to achieve. All of them, no matter how well intentioned, add a compliance burden and legal uncertainty to an industry that is both vital to the national interest and has yet to fully find its footing.

Having a single federal regulatory framework would not allow the industry to run amok. Far from it; most industries regulated by the federal government do not have an easy or trivial time complying. However, what they do have is the certainty that a single set of standards, carefully complied with, will allow them to do business throughout the United States.

It is also not reasonable for consumers and businesses in one state to be beholden to the combined decisions of legislators across the country; far from providing a laboratory of democracy where each state may choose for itself, most of the proposed regulatory burdens would be inescapable and in effect national. Instead of giving each state the ability to pick its own way of handling AI, the bulk of the regulations would bury companies under the joint weight of the legislation of all the states together.

We urge Congress to pass a common-sense national AI framework, not one that gives the industry a free pass, but one that will give companies and consumers from across the country a measure of certainty as they conduct business and move forward together.