President Biden is establishing the United States Artificial Intelligence Safety Institute. The aim is to bring America in line with Communist China and other nations. The decision raises concerns about whether government control of the emerging technology is any less a threat than AI itself.

In a statement, the Department of Commerce wrote that it’s establishing the new bureaucracy “to lead the U.S. government’s efforts on AI safety and trust, particularly for evaluating the most advanced AI models.”

The announcement implemented an executive order that Mr. Biden issued days earlier. The commerce secretary, Gina Raimondo, wrote that her department “plays a pivotal role” in the government’s approach to “seizing the potential that comes with the development of advanced AI, while mitigating dangerous capabilities or risks to safety.”

Senator Warner of Virgina, a Democrat, said at Politico’s Global Tech Summit that America is “in an enormous technology competition, particularly with China, and that national security means winning the battle around AI.”

Beijing, Mr. Warner said, “is very much ahead of the game in terms of self-regulating AI.” A Republican of Texas, Senator Cruz, said at the event that it “would be profoundly dangerous to the United States, from a national defense perspective, but also certainly from an economic perspective” if Communist China masters AI first.

Lost in the debate is faith in America’s constitutional guarantees of liberty. Communist states, after all, can always be counted on to be behind those that encourage entrepreneurism. Emulating their approach of top-down control, far from encouraging innovation, only suppresses it.

Last week, America joined 27 other nations at Britain’s international summit on AI to back the Bletchley Declaration, a pledge to cooperate in ensuring AI is “human-centric, trustworthy, and responsible.” None are words that apply to Communist China.

The summit’s goal, as the Hill describes it, is to “quell fears of the technology leading to human extinction.” Such dramatic language invokes science fiction dystopias, obscuring a more familiar danger: governments grabbing power at the expense of individual liberty.

As long as there have been governments, they’ve used threats real and imagined to accumulate power. It’s easy to do, as human beings are inclined to fear new machines. These “technopanics” have occurred for tools as quaint as the printing press.

In 1096, Pope Urban II banned a fearsome weapon, first invented in Europe by the Greeks, then gaining popularity: the crossbow. He threatened to excommunicate any Catholic who used it against fellow Christians, earning them an afterlife sentence in hell.

An abbot in the 15th century, Johannes Trithemius, worried that machines rolling off copies of documents would steal the livelihood of monks. If they no longer had to copy text by hand, wouldn’t they fall prey to the deadly sin of sloth?

Closer to our own time, the press and government agencies were full of warnings about the evils lurking on the World Wide Web. While there have been new digital dangers, for the most part they have been old crimes committed with an assist from new technology.

In a conversation with the prime minister of the U.K., Rishi Sunak, entrepreneur Elon Musk said the role of government in AI technology should be that of a referee in a sports match. A referee, though, has no stake in who wins. The same cannot be said of governments.

Digital superintelligence, Mr. Musk said, “does pose a risk,” and so “there is a role for government to play to safeguard the interests of the public.” He agreed with concerns in Silicon Valley that regulations will be “annoying,” but approved of them.

More troubling, Mr. Musk took Communist China at its word that it will abide by international agreements, though it tends to ignore those that run contrary to its interests on everything from intellectual property to maritime borders.

If America’s goal is to harness AI’s potential, the wise path is to keep the hand of government out of it as much as possible, freeing inventive minds to deliver a technological edge. Like the crossbow, AI is neither inherently good nor bad. It just depends on whose finger is on the trigger.