I have spent half a century tracking threats to American security — from Soviet armored columns to jihadist networks to the accelerating military machine of communist China. In all that time, I have never watched a danger develop quite like this one: a threat born not only from our adversaries, but one we are building ourselves, with our own capital and our own engineering genius, moving faster than we have decided what rules should govern it.
That is the unspoken risk inside this technology race. Right now, America is moving too fast to see it clearly.
To understand where this road leads without discipline, look at what Beijing has already built. In a 2025 address to China’s Politburo, President Xi Jinping called machine intelligence a “strategic technology” reshaping the foundations of state power — not merely a tool, but the engine of governance and global dominance. China has deployed more than 200 million surveillance cameras, many equipped with facial recognition and integrated into national police networks.
Human Rights Watch documented that a digital surveillance network in Xinjiang flagged Uyghur Muslims for detention not because they committed crimes, but because an AI-powered predictive surveillance system said they might.
China now holds roughly 70% of global surveillance patents, and through the Belt and Road Initiative it is exporting that model of control across Asia, Africa and Latin America. Countries importing these systems are not buying hardware. They are importing a governing philosophy — one where automated control has displaced constitutional rights. That is the road we must not walk. And right now, we are not as far from it as Americans assume.
America is racing to answer. It should be. On his first full day in office, President Donald Trump announced the Stargate Project — committing OpenAI, Oracle, SoftBank and MGX to up to $500 billion in U.S. computing infrastructure, with the flagship campus in Abilene, Texas, already operational.
The administration’s July 2025 AI Action Plan outlined more than 90 policy actions spanning innovation, infrastructure and international leadership. In September, Trump convened more than 30 technology executives at the White House — Apple, Google, Meta, Microsoft, Oracle, AMD and OpenAI — securing investment pledges topping $1.5 trillion through the decade.
In November, the president signed the Genesis Mission executive order, establishing the most ambitious federal research initiative since the Manhattan Project: a national platform fusing Department of Energy supercomputers, secure cloud systems and scientific datasets to compress discovery cycles from years to months.
Microsoft just committed $10 billion to technology infrastructure in Japan, anchoring a U.S.-aligned digital ecosystem in the Pacific as a direct counter to Beijing. The pace and scale of this investment is right. The urgency is warranted.
But urgency is precisely when guardrails disappear. And that is why the alarm must be sounded.
AI RAISES THE STAKES FOR NATIONAL SECURITY. HERE’S HOW TO GET IT RIGHT
The same systems built to outcompete Beijing can be turned inward — not by any single dramatic decree from the White House, but through thousands of small decisions made in the name of speed: efficiency replacing accountability, automation replacing human judgment, convenience replacing constitutional limits.
Cybersecurity experts now warn that autonomous systems resist reliable control through conventional software frameworks. That is a sobering reality when those systems are embedded in defense, law enforcement, or the delivery of government services.
Freedom in America is rarely lost all at once. It erodes through systems that make decisions too fast to question, operate too opaquely to challenge, and reach too widely to escape. When automated systems begin to determine who receives benefits, shape what information citizens can access, or drive consequential decisions without human accountability, authority has quietly migrated — from elected officials and courts to systems no one fully understands and no one voted to empower.
Trump’s December 2025 executive order on AI rightly pushed back against the splintered maze of state regulations threatening to fragment American innovation. His March 2026 national AI legislative framework urged Congress toward a unified federal approach covering child safety, intellectual property, free speech and workforce development. Both were necessary steps. But a legislative recommendation is not law.
CLICK HERE FOR MORE FOX NEWS OPINION
There are still no binding federal standards specifying what human oversight is required before an automated system makes a binding decision about an American citizen, what transparency is owed when government systems evaluate the people they govern, or how privacy, due process, and free expression survive the age of machine rule. Congress must act — not to slow the race, but to make sure what we preserve is worth winning.
These questions are at the core of my new book, “The New AI Cold War: Liberty vs. Tyranny in the Age of Machine Empires,” available later in April. The contest with China is as real as any this nation has faced. So is the internal temptation every great power confronts in a long competition: adopting the logic of your adversary in the name of defeating him.
Make no mistake — our adversaries are already deploying these technologies against American interests. But that is not where the greatest danger lies. It lies closer to home: that in the name of defeating them, we quietly build the same architecture of control ourselves — and by the time we recognize it, the infrastructure is already in place.
History will not judge us by whether we built AI first. It will judge whether we remained free while doing so.
