Almost overnight, the impact and potential threats posed by artificial intelligence have become one of the biggest issues in America. The pre-election dialogue revolves around two very serious questions. First, will rogue or malevolently deployed AI exterminate humanity, and second, will the data centers needed to develop AI overwhelm resources and destroy quality of life wherever they’re constructed?
Read more Scores on PISA Are Leaning the Wrong Way
The consequences of how these questions are answered cannot be overstated. But the debate is almost moot, because if the United States doesn’t develop AI and the data centers required to support them as fast as it can, it will not be able to deter World War III. Period. That, too, is an existential threat.
Almost every foreign policy and economic goal of the Trump administration recognizes a handful of simple, painful truths. China is a fascist, expansionist superpower. Its allies include Russia, North Korea, and Iran. And for the last several decades, Chinese agents have been stealing every leading-edge technology from the U.S. they can get their hands on; they have built a government-subsidized manufacturing base that dwarfs that of any other nation, and they are determined to surpass the United States in AI technology.
The military value of AI is not incremental. It has completely transformed warfare and has left legacy platforms obsolete. Without fundamental modifications and continuous innovation, multi-billion-dollar assets, from aircraft carriers to ICBMs, are now obsolete. The pace at which AI capacities advance means that even if known new AI-enabled threats are addressed—for example, low-cost autonomous drones that can be released by the thousands in coordinated and precisely targeted swarms, or swarms of virtual AI agents launching cyberattacks against critical infrastructure—by the time countermeasures are developed, there will be a new generation of even more potent offensive AI technology.
None of this should be news. We are fighting another Cold War, with AI replacing ICBMs as the instrument of mutually assured destruction. But on the eve of a midterm election, politicians in both major parties, joined by pundits across the entire ideological spectrum of media properties, are propelling a populist wave of anti-AI sentiments.
People tend to forget how fast the world can slide from relative peace to total war. American actions affecting Venezuela, Greenland, Cuba, the Panama Canal, Diego Garcia, the Strait of Malacca, most of South America, and, of course, Iran, are all steps designed to avoid the predicament America faced in 1941. In the run-up to the catastrophic conflict of World War II, decades of disarmament and appeasement had enabled the Axis powers to acquire significant territory and resources. What might have been stopped with far less bloodshed a few years earlier turned instead into the most devastating war the world had ever seen.
China now is Germany then. Russia is still Russia. And Iran is Spain in 1936.
We have no choice. America has to build AI as fast as it can, and it has to dominate the technology. The only thing we can reasonably expect to control is how we do it.
Read more Trump Blasts ‘Globalist’ ICC, UNESCO, and Human Rights Council in UN Address
In this, we have many options. Some of the CEOs of the biggest companies developing AI have called for a slowdown and more government oversight. This is a reasonable proposition, but it will undermine one of the biggest competitive advantages America has against China, which is our competitive startup culture. For all we know, it could be an AI startup, and there are thousands of them now in America that build the tool able to stop a runaway AI swarm that has escaped from one of the major players.
Imagine if AI ends up heavily regulated, and starting in 2029 another Obama/Biden-type administration occupies the White House. Imagine that every iteration of new AI capacity has to be approved by bureaucrats chosen for their ideological reliability and their identity as members of underrepresented groups. Even if competent bureaucrats were the gatekeepers, the inherent speed of AI advancement would be fatally compromised. Perhaps some oversight is called for. But watch out. We can lose this race to China, and if so, all the caution in the world won’t save us from what comes next.
To the other question, which is how data centers will be deployed, there is some good news. Recent innovations in closed-loop cooling technologies mean it is now possible to operate data centers with zero operating losses of water. Once they are primed, they will not consume any water. And as processor efficiency continues to improve, less electricity will be required. Eventually, data centers may be in orbit, where access to perpetual sunlight and the cooling properties of outer space will make their operations even more efficient.
With respect to power consumption, in any case, the fact that technology companies are demanding more electricity from the U.S. grid is a healthy political development. Instead of a totally unrealistic, quasi-suicidal energy rationing mentality, something that was dominant in the Obama/Biden era, we have lobbyists representing the most influential economic sector in America telling our politicians it’s time to build more electrical generating capacity. Good.
The footprint of planned data centers is indeed on a scale that is almost unbelievably huge. And because data centers require human labor to build and operate, they can’t be built in the middle of nowhere. They have to be within commuting distance of population centers. But communities can require mitigating standards that tech companies can easily afford. They have to be energy- and water-neutral, if not positive. And they have to be sited in a way that concentrates them into areas with the least impact. Moreover, the sound they make and their visual impact have to be mitigated by landscaped earth berms surrounding the data centers, and those berms have to be as high and as wide as are needed to completely shield them from surrounding properties.
America has no choice. Artificial intelligence is here to stay, and the risks to our future are far greater if we lose control of its direction. We must build it and stay one generation ahead of our enemies. And while it, in that way, will deter the next big war, it may also bring us wondrous and beneficial scientific breakthroughs, helping us cure diseases, eliminate poverty, and heal the planet.
