Geniuses are an odd species. To paraphrase Montesquieu, the exceptional are often also a bit loony. Cornelius Vanderbilt, a devotee of spiritualism, believed himself to have communicated with the spirit of George Washington. To improve his health, Henry Ford devoured weeds, and Thomas Alva Edison was averse to bathing. The point: geniuses are an odd species, who, for all the indispensability of their works to American prosperity and innovation, ought not be trusted credulously. Today in Washington, D.C., Congress must endeavor to protect artificial intelligence (AI) from the technologists most prominent in its creation.
Anthropic’s Dario Amodei desires the U.S. government (an agent of compulsion) to dictate the “pace” of innovation, as the CEO wrote this month. “I have become convinced that fully addressing the risks requires even more prudence—not just investing in risk prevention, but pacing the rate of capabilities advancement so that risk prevention has time to keep up.” Amodei warns of impending catastrophe should innovation continue at its current rate. Yet his company, the creator of what he supposes to be a tool of catastrophic potential, demands state-imposed fetters. Anthropic will institute some precautions, but not thoroughgoing “safety standards as well as limits on the rate of unchecked AI progress”—not until such standards and limits bind the entire industry. This means, in simpler terms, the kind of regulatory scheme Anthropic prefers—the kind that jars with the principles of antitrust law and would, as David Sacks notes, shelter the company’s market share from the gales of competition.
The “prudence” of Amodei and the other pessimists is more rascally than genuine; it is an age-old propensity to see evil lurking in every major innovation and demand statism to manage it. The experience of the ages did not vindicate such assumptions when the printing press brought on a flood of “confusing and harmful” literature, when the typewriter displaced the human hand, or when the emergence of machines caused economist John Maynard Keynes, in the 1930s, to declare: “We are being afflicted with a new disease, technological unemployment.”
In 2016, Geoffrey Hinton, the “Godfather of AI,” predicted the demise of human radiologists, which field is, a decade on, employing human beings at record rates. Explaining his error, Hinton recently confessed that, a decade ago, he was ignorant of important facts. Unreformed, Hinton recently likened an episode in which an OpenAI agent, improperly set free from technical constraints, hacked Hugging Face to the nuclear reactor explosion at Chernobyl. He allowed Congress about a year to fend off disaster by regulating the technology.
The regulatory model favored by the pessimists is likely to subject free innovation to cartelization, free speech to state control, and the American system of property rights and free exchange to the will of bureaucratic juntas. As the Taxpayers Protection Alliance argued, even those pessimists who “consider themselves conservatives on technological questions…propose to embark upon a revolution in American government. To preserve the good life, they tell us, we must jettison the freedoms on which the pursuit of happiness depends. No vain denials of the scope of proposed state control can convincingly gainsay that contradiction.”
The marriage of pessimism and statism has arrived on Capitol Hill, where many are in a mood to regulate AI—and to do so aggressively. Myriad bills have been introduced, and—among others, presumably—Senate Majority Leader John Thune (R-S.D.) and Sen. Amy Klobuchar (D-Minn.) are concocting what is likely to be a prominent addition to the collection. House Minority Leader Hakeem Jeffries (D-N.Y.) recently articulated the fearful position well, advocating “decisive congressional action immediately, in a manner consistent with what some of the leading AI voices in the country are now saying needs to happen.”
All the while, President Donald Trump and many in Congress—including Speaker Mike Johnson (R-La.)—have resisted the paranoiacs. They remain confident in the American system: free innovation within the confines of traditional legal principles developed throughout the course of centuries to protect the rights of citizens from corporate excesses.
Doubtless, the captains of industry now building frontier AI models can accomplish innovative feats beyond the ken and capacities of all but a few Americans, let alone the median Washington, D.C., politician or bureaucrat. But their knowledge of politics and economics—of the ends of government and the means necessary to secure them—is not to be assumed. It is the office of the technologist to understand how to innovate, and that of the congressman to understand how to govern. The fretting and the myopias of the one should not rule the judgment of the other. A federal regulatory standard, enacted by Congress, is a sine qua non of sustained American innovation. But it must be built upon a foundation of practicality and not fanciful timidity, experience and not theory, and sound principles—the principles native to this land of liberty—and not statist reaction.
