Is artificial intelligence a blessing or a curse? It may be one of the defining questions of the 21st century. AI promises extraordinary benefits, from medical breakthroughs to unprecedented economic growth. But its development also carries profound risks for the financial and geopolitical landscape—and shows no signs of slowing down. On Monday, Anthropic filed plans for an initial public offering, a move that could value the company at nearly $1 trillion. Other major AI companies are already looking to follow suit.
Against this backdrop, President Trump on Tuesday signed an executive order asking some AI companies to allow the government to review powerful new models 30 days before releasing them to the public—a scaled-back version of an order originally scheduled for last month.
Today, historian and Free Press columnist Niall Ferguson argues that for all its potential for good, the rapid and largely unregulated growth of AI doesn’t just threaten to render us obsolete; it has sparked the most dangerous arms race in history between the U.S. and China. And, like the nuclear arms race before it, he warns, the dangers it poses will not expire of their own volition. —The Editors
In 1957, Henry Kissinger published Nuclear Weapons and Foreign Policy. The book clearly identified the central problem of the nuclear age: that any strategy of brinkmanship based on the threat to use strategic nuclear weapons created such a high risk of Armageddon that it lacked credibility. Though flawed in the eyes of its many critics, it was a seminal work that established his reputation as a strategic thinker at a time of public uncertainty as the Cold War escalated.
Seven decades later, at another moment of uncertainty and escalation, we desperately need someone like Kissinger to write Artificial Intelligence and Global Security.
Ends today: The Free Press Community Sale
New subscribers to The Free Press get 20 percent off their first year.
Think for yourself, not by yourself.
The unfolding history of artificial intelligence has now arrived at what may be its most dangerous moment. There are two barely controlled AI races, one between around five American companies—Anthropic, Google, Meta, OpenAI, and xAI lead the field—and the other between the two geopolitical superpowers: the United States and China, with its own competing companies.
The leadership of the competitors in this race is, to say the least, of mixed quality. The chief executives of the most important companies include at least one with a record of duplicity, and at least two egomaniacs. Meanwhile, the president of the United States is a former real estate developer and reality TV star, roughly half of whose public utterances are mere bluffs, and the leader of the People’s Republic of China is a Marxist-Leninist who aspires to eclipse Mao Zedong as a dictator.
These two AI races have several inherent constraints, including the amount of data available for training models, the supply of computational power, and the capacity of capital markets to finance trillions of dollars in capital expenditures.
But the most important constraint—regulation—does not feature on that list. And because of that, we may be hurtling toward the most dangerous arms race in history.
