Dean Ball, board member of The Alexander Hamilton Institute for the Study of Western Civilization (AHI), is taking his in-depth analyses of artificial intelligence to a new level by writing at greater length on how society should respond to and would most fully benefit from its mounting progress. He has been working on a long essay, which might even be a short book, called “Artificial Intelligence and American Rebirth” that deals with such themes. In a piece at his online newsletter Hyperdimensional, he shares some of his thoughts.

One is that AI and its effects are, or will be, simply too large and diffuse—comparable in this sense to the Industrial Revolution—to be governed from the top down. “There is no way to pass ‘a law,’ or a set of laws, to control an industrial revolution,” writes Mr. Ball, a research fellow at George Mason University’s Mercatus Center and a former key staff member of the Hoover Institution.

His new piece, “Here’s What I Think We Should Do,” advocates instead a major emphasis on ensuring a full, accurate understanding of various AI systems’ capabilities. Government should get innovative AI labs to adhere to “a basic standard of transparency.”

An “alumnus” of the AHI who was one of its undergraduate fellows as a Hamilton College student in the early 2010s, Mr. Ball also worries that a major stifling AI, in hopes of thoroughly minimizing its risks, is already underway:  “A large bureaucratic apparatus has begun to emerge around [the] nebulous and all-encompassing conception of AI safety. This apparatus should be rethought.” A major part of it is the National Institute for Standards and Technology’s AI Risk Management Framework (RMF).

In current form, RMF is “a comically overbroad document, aiming to present a fully general framework for mitigating all risks of all kinds against all people, organizations, and even ‘the environment.’” Unreasonably, it advises developers and corporate users of AI to take nearly every conceivable step to mitigate risk. RMF is “quickly becoming a de facto law, with state legislation imposing it as a minimum standard,” while advocates urge the Federal Trade Commission to enforce it nationally.

Treating it as a law, Mr. Ball warns, would impose an AI legal framework comparable to the National Environmental Policy Act of 1969, “creating an opportunity for anyone to sue any developer or corporate user of AI for, effectively, anything.” To prevent this, he argues for jettisoning RMF and supplanting it with “a far more focused document.”

Mr. Ball also sees national security concerns and risks for America’s economic competitiveness that will arise from movement toward major state-level regulation. Some states are “barreling ahead with … laws that create a dense thicket of open-ended rules for both developers and corporate users of AI to follow.”  One such law has already been enacted in Colorado.

Lacking insight into the “geopolitical and national security dynamics of AI development,” state governments shouldn’t be “making decisions … for the entire country” about its trajectory—one reason federal laws should try to prevent much, although not all, state-level regulation of AI.

Mr. Ball suggests federal preemption of potential state laws in, for example, the area of liability. He would like to see a nationwide legal-liability standard—“rooted in the basic American concept of personal responsibility: a rebuttable presumption of user responsibility … when someone misuses a model, the law [would presume] they are responsible unless they can demonstrate that the model ‘misbehaved’ in some way that they could not have reasonably anticipated.”

Much of the emphasis in artificial-intelligence law should be on ensuring “a high-quality information environment.” If a doctor, for instance, “wants to know whether his use of an AI model … is a medical malpractice risk, his inquiry will not be meaningfully aided by a law that says, ‘do not use AI models irresponsibly.’ Instead, he will need high-quality evaluations to gauge whether different models are appropriate.”

Mr. Ball notes, too, that AI depends and will depend upon a “vast industrial infrastructure.” Thus, reform of the process for building permits should perhaps be the government’s “most important AI policy priority … It is critical to ensure the construction of the energy generation, data centers, semiconductor manufacturing facilities, and other physical infrastructure we will require.”