The Pentagon’s Dangerous Rush to Adopt and Deploy AI into All Military Missions

Interview with William D. Hartung, senior research fellow, Quincy Institute for Responsible Statecraft, conducted by Scott Harris

William Hartung discusses his recent Nation magazine article, “The Pentagon is Going ‘AI First,’” and issues related to growing concern about the unregulated military applications of AI. Hartung is co-author with Ben Freeman of The Trillion Dollar War Machine: How Runaway Military Spending Drives America Into Foreign Wars and Bankrupts Us At Home.

SCOTT HARRIS: We begin our program tonight by welcoming back to our show William Hartung, senior research fellow at the Quincy Institute for Responsible Statecraft and author with Ben Freeman of the book, The Trillion Dollar War Machine: How Runaway Military Spending Drives America into Foreign Wars and Bankrupts Us at Home. William, thank you so much for making time to be on our program this evening.
WILLIAM HARTUNG: Yeah, it’s great to be back. It’s been a little while and things haven’t been getting better.
SCOTT HARRIS: (Laughs) I laugh, but yeah, it’s very serious, what we’re going through as a country and the world. So your recent article I think is so important and we’re going to get to that in a moment. “The Pentagon is Going AI First,” written by yourself and Janet Abou-Elias, both of Quincy Institute. And I want to get to artificial intelligence and its applications in the military and big concerns about that. But I thought because you’ve been certainly watching the news about this peace agreement, what’s your view of this memo of understanding between the U.S. and Iran? That could be signed as early as this Friday. People are skeptical.
WILLIAM HARTUNG: Well, if it stops the killing even for 60 days, that’s a good thing. Now, how exactly Trump is going to deliver Netanyahu, who seems to have no intention of stopping the attacks on Lebanon? I don’t know. Will Iran sort of fudge it because they want to move to the next step in peace talks or will this upset the whole thing? We don’t know. But the thing that galls me the most is Trump launched an illegal war. They bombed an elementary school. There were nurses running out of hospitals with babies in incubators. They let Israel go ahead and occupy and do ethnic cleansing, displace a million people in Lebanon with U.S. money and guns. 

And we’re supposed to believe that Donald Trump is a peacemaker. And the mainstream media, intentionally or not, “Trump Reaches Peace Deal or Peace Accord or Ceasefire.” Well, what about “Trump Disrupts the Entire Middle East and the Global Economy for No Good Reason and Then Saves His Ass?”

I mean, that would be a good headline.

SCOTT HARRIS: Definitely. Yeah. No, there’s not enough skepticism. And one more comment, if you would. I am just constantly amazed at how Trump’s very transparent manipulation of the stock market has succeeded beyond his wildest dreams. He says three or four words about a peace agreement or energy prices and they react like salivating dogs, you know. It’s wild.

WILLIAM HARTUNG: Oh yeah. Well, the markets are like that. I mean, it’s one of the kind of dysfunctional features of capitalism. I mean, they can’t see beyond their nose. But the other thing is the polymarkets. Apparently I haven’t seen this written in blood, but it’s believed that people in the Trump administration are betting on the poly markets like, “Are we going to bomb Iran on Sunday?” Of course, they would know.

So it’s quite possible they’re profiting immediately. They’re not waiting after their career to cash in, go work for a contractor. It’s the most obscene thing imaginable. Sen. (Jeff) Merkley of Oregon has a bill on this, which is languishing, which would ban (government officials) from being in those (prediction) markets. But it’s an addiction. I mean, there’s people in the nonprofit sector, maybe we should have that for us. We know about foreign policy. We shouldn’t be betting on this stuff.

And people are like, “No, no, I love the poly markets.” It’s like sports betting except people get killed. It’s terrible.

SCOTT HARRIS: Yeah. I read about horrific things like people in Israel betting on Iranian missiles, whether they were going to hit or be intercepted. I mean, it’s frightening. Well, William, go ahead. If you wanted to comment on that, go ahead.

WILLIAM HARTUNG: Oh, well, it’s all of piece with the rise of the tech bros because they’re trying to remove us further and further from our humanity and get us more and more into being addicted to machines and betting and casinos and games, all of which have a place. But not if you lose your way as a human being, which seems to be their goal.
SCOTT HARRIS: Right. And this sort of ties in with our topic tonight. Your article, the “Pentagon is Going AI First,” again written with Janet Abou-Elias, both you and her from the Quincy Institute. You discussed the Pentagon under Donald Trump and defense secretary, aka “Secretary of War Crimes” Pete Hegseth, that they’re going full speed on employing AI systems across all of the U.S. military’s missions with little or no regulations or controls. I wondered if you just provide an overview of what exactly is going on in the Pentagon with AI?

WILLIAM HARTUNG: Well, they’re probably as usual over-assessing what this technology can actually do, but they want to do everything from process paperwork to better targeting systems to … There’s this idea not—just drones, pilotless aircraft—but uncrewed ships and armored vehicles that don’t have anybody in them. And of course, Silicon Valley put a big bet down on Donald Trump after they brought in JD Vance, who’s a product of Silicon Valley as the VP. That’s when Elon Musk came in with the $250 million to try to get Trump elected. That’s when they made Vance their man to go to Silicon Valley, get even more money. So they were thrilled at Vance’s appointment because it was like they literally had their guy in the White House. They’ve also got people in the Pentagon and many of the other agencies. And of course when they let Musk and his people dismantle large parts of the government, they sent a bunch of 25-year-old tech bros in to basically take apart agencies they didn’t understand.

So they’ve got a huge purchase on Donald Trump. He had his crypto ball and they’ve all put in Don Jr.’s on the board of a drone company and UAE’s putting money in his crypto fund. And before that the Saudis put $2 billion in Jared’s investment fund in hopes that Trump would come back, which he did.

The corruption is probably unprecedented in our history, especially because it’s happening so quickly and there’s so many corrupt actors and also half the time Trump brags about it. I mean, he’s not hiding it.

SCOTT HARRIS: No. That’s true.

WILLIAM HARTUNG: He even has a piece of the UFC that did a little stunt on the White House lawn.
SCOTT HARRIS: What was the stunt relating to—

WILLIAM HARTUNG: Well, they had the ultimate fighting match for his birthday.
SCOTT HARRIS: Right. No, yeah.
WILLIAM HARTUNG: They did a big glitzy thing about it. Well, Trump has stock in the Ultimate Fighting, UFC promotion business. So if that draws more people to watch the stuff, his stock goes up. So you can’t find anything related to Trump that doesn’t somehow involve enriching Trump or his merry band of grifters, you know.
SCOTT HARRIS: Right. The endless dimensions of corruption and I think, rightly called. And we’ll probably know more in a couple of years, but rightly called the most corrupt presidential administration in U.S. history. I think that’ll be borne out because it’s all being done in the open, as you said.
WILLIAM HARTUNG: It’s like the Teapot Dome Scandal, but it was one scandal. This is the scandal by the day, by the hour, by the minute, you know.
SCOTT HARRIS: Yeah. So what are the dangers in your view of having few, if any, controls on artificial intelligence being integrated into military applications? And I would just mention that I know very little about this topic, but I do know there was some reporting about the Israeli military using artificial intelligence in what many people describe as a genocide in Gaza, where more than 73,000 people have been killed there, many women and children. Civilian targets were obliterated on almost a daily basis. And even since the ceasefire, people are still being killed regularly in Gaza. I also have certainly read about concerns that AI might have been used by the Pentagon in the first day of the U.S.-Israeli war against Iran where a girls’ elementary school was obliterated by a U.S. Tomahawk missile, killing 145 to 175 girls and their teachers.
WILLIAM HARTUNG: Yeah. Well, and Israel Palantir, which is run by Peter Thiel. Palantir’s Peter Thiel was the one who employed JD Vance and financed his Senate campaign. He’s Peter Thiel’s guy. They actually, the president of Palantir held their board meeting in Israel during the Gaza War to cheerlead for the war. It wasn’t enough to cash the checks. He wanted to be front and center. And there was a thing called Project Lavender, which basically was a software targeting system that helped them generate more targets more quickly. So basically they were hitting more targets, doing more destruction in a given amount of time than they would have done without that technology. And back when the Saudis were devastating Yemen, they said, “Oh, you want targeting, you want accuracy, you reduce civilian casualties—only if that’s your goal. If your goal is collective punishment and killing as many people as possible, AI makes that easier to do.
And there’s some dispute about the elementary school. Was it AI-controlled, that attack or not? But the point is these things, if they work and they’re willed to do with ill intent, they’ll devastate people. If they don’t work, they could cost accidental slaughter of people. And if the Pentagon and the government thinks they can wage war where they don’t put a single U.S. military person at risk, they’ll fight more wars. And of course they think that devastation equals domination, but all it does is kick the struggle down the road and people use whatever means they have to fight back. So they’re not going to fight back with AI, but it could be terrorism, it could be disrupting trade, it could be … The idea that you can just bully and destroy a country, kill its leaders and its people and they’re going to sit on their hands is the most naive thing I’ve ever heard in my life.
And that’s what this war with Iran has been like. That’s what Israel’s genocide in Gaza has been like. That’s what the occupation and ethnic cleansing in Lebanon is like. They think if they do enough destruction, enough intimidation, enough bullying, people are just going to roll over and say, “I salute you.” Well, that’s not happening. It’s not going to happen. The problem is they’ve got this sort of system that’s really ultimately going to undermine the power of the U.S. and the world, but in the meantime, the destruction is moving rapidly and the ability to organize against it takes somewhat longer. So we’re kind of in a race against time to reassert some sort of rules. I mean, even the Biden folks, like how can they just talk with a straight face about a rules-based international order when they not only turned a blind eye to the genocide in Gaza, but they help finance it and back it politically by protecting Israel at the U.N.
So it’s kind of really a pivotal moment in the history of this country and this world. And I think some people are waking up to that and some people just wish it would go away.
SCOTT HARRIS: Let me reintroduce you to our audience. We’re speaking this evening on Counterpoint with William Hartung, senior research fellow at the Quincy Institute for Responsible Statecraft. And we’re talking about his recent article co-written with Janet Abou-Elias titled, “The Pentagon is Going AI First.” And William, in the article, you discuss concern about the growing power of private for-profit corporations with research or AI research funded by venture capital firms. In fact, you talked about a $5.6 billion 10 year-contract that’s been signed by the Pentagon with a company called Salesforce. Talk about the concern we have about the for-profit companies tied in with AI technology.

WILLIAM HARTUNG: Well, it’s amazing because Trump, if he had his way, would spend $1.5 trillion through the Pentagon next year. He probably won’t get all of it, but that’s the highest level since World War II. But then on top of that, you have venture capital firms putting private money into military tech startups. So that’s more money out of the economy going to the military sector.

And their idea is if they invest in 10 companies and one of them hits, they’ll get rich and they have this thing called the Valley of Death. How do you get from being a small tech company to getting Pentagon contracts? And their answer is you simplify the bidding process, you rig it, you put your people in government, you get rid of all kinds of regulations about “Does this thing work? Are they price gouging us?” You just basically create a system where you throw money at these tech companies and they say, “Trust us.” So that’s sort of the stage that we’re at and they’re very rhetorically at least, they look down on the old companies like Lockheed Martin, which of course has some purchase in Connecticut because they’re like, “Oh, they’re slow, they’re cumbersome, they don’t know the software, they don’t have the business model.” They should just get a participation trophy and go home.

But beneath that, they also do partner with them as subcontractors and on some of these things like Golden Dome, the impossible dream of this flawless missile defense that Trump is so fond of. The hardware will likely be made by the Raytheons and Lockheed Martins of the world and the software and the targeting and the systems integration will be done by firms like Palantir and Anduril (Industries) run by this 32-year-old gamer named Palmer Luckey and SpaceX and Elon Musk. So if they drive the budget high enough, they can keep both groups happy, the old guard and the new guard. If we actually put limits on the Pentagon budget, which would be a great idea, they might fight it out. At least they’d do some damage to each other’s reputations, some information would emerge. But at the rate they’re going, it’s more like a partnership or sort of like, well, it’s like fake wrestling, which Trump loves.

I mean, they’ll have rhetorical barbs, but they’ll laugh all the way to the bank.

Yeah, it’s really something that … And there’s been resistance in Silicon Valley. There was resistance to Google getting involved in drone targeting. There’s a No Tech for Apartheid organization. But people have been blacklisted in Silicon Valley for coming out against military work and so they’re kind of regrouping. And some of the smaller firms would rather be doing things like environmental research, but one of them was going to use their technology to track the state of the polar ice caps. And this VC firm said, “Oh God, no. I mean, that stuff would be as amazing military applications. Let’s give you money to do that. ” So the VC firms are shifting the balance in Silicon Valley in a more hawkish direction and they’re among the most vocal, not only about how the U.S. can dominate the world with the right technology, which has never happened and never will happen, but also they fancy themselves as some sort of techno gods.

We’re supposed to thank them for this. They’re saving us and by the way, we’re making Musk a trillonaire along the way.

It’s completely unhinged, which I said that about Palmer Luckey from Anduril and he started attacking me on X because he said I was a shill. And I said, “Well, who am I shilling for? The peace movement? What are you talking about? ” But anyway, the fact that he has time to do that is a bit scary, too, because he’s running this big company and he’s busy getting offended by articles in Tom Dispatch. So that’s the story unto itself. But anyway, yeah, that whole new surge of private money and I would say also that the tech sector is a lot more vocal about their prejudices and like Palmer Luckey goes on 60 Minutes and says, “We’re going to bury China” or “Palantir celebrates genocide in Gaza.” The head of Lockheed Martin would be more circumspect. He would like on the investor colleagues say, “Well, that turbulence in the Middle East will create a market for our products so we can help our allies promote …” They would try to cover it up with antiseptic language and the tech folks don’t do that.

I mean, it’s almost as bad as like Alice’s Restaurant, where they probably say, “I want to kill. I want to kill.” It’s not quite that bad, but it’s kind of their vibe.

SCOTT HARRIS: Group W Bench, yes. I did want to ask you about maybe one sign of hope among all these tech companies. Anthropic, the creator of the Claude AI has resisted the Pentagon’s use of some their software and they were blacklisted for it. What can you tell us about Anthropic and what it could mean in terms of maybe some centers of resistance for these autonomous weapon systems that are likely to violate international law, the Geneva Conventions and as you said very clearly in your article, it’s civilians in these conflict zones that will suffer and die.

WILLIAM HARTUNG: Well, I think we’re going to need resistance from inside the tech sector. So in that sense, Anthropic, that stand is important. The problem is they wanted to sell stuff to the Pentagon with conditions of how it could be used. But once you give stuff to the Pentagon, your ability to control how they use it is very limited. So then they would have to decide just not to sell to the Pentagon.

And is there a big enough market doing non-military stuff for them to do that? There’s also a little wrinkle. I don’t know that Anthropic is a culprit here, but some of the tech firms want just enough regulation to keep smaller companies from getting in the game. So this discussion of regulation really needs … We need to know what they mean by that because I mean, there’s been bills about what kind of data they can collect on us, so that’s important. But if they can get boatloads of military money for things that aren’t even properly tested in a doctrine that says we’re going to use this technology to dominate the world, if you don’t stand up to all of that, then some of these smaller issues of regulation are going to get just blasted out of the water or cast aside.

So I think Anthropic, I think the workers movements that have come and gone about not doing this kind of work, I think there’s people in this sector who want to use AI for positive purposes. I think we’ve got to be a litle careful about that because even the so-called positive elements have their detriments. What if AI wrote everything for us, Shakespeare, whatever we wanted. Our skills with atrophy, that’s not a good thing. Or Elon Musk is, “Well, it may eliminate jobs where you can have guaranteed basic income.” A lot of people, work is tough, but a lot of people do build their identity around their work. If you have no job and no prospects and they give you a check once a month, what kind of life is that if it’s being run by the tech folks? So I think there’s a lot of issues, but I think there might be some positive elements in the use of some of this technology if it’s carefully monitored. But are those positive elements going to make you as much money as going all in on war?

So then it really comes down to, do you want to survive as a civilization and what kind of world we want to live in? What kind of human beings we want to be? And so it becomes very fundamental.

SCOTT HARRIS: Huge issues there. We only have a couple minutes left, but I did want to ask you one quick question that deserves an hour in itself. But as you said, there’s a lot of concern about people losing their jobs through the employment of AI across the board. And then as we’ve been talking about, you’ve got the dangers of AI being employed in military applications. What are some of your recommendations for how our country and the world should put in place some regulatory standards and controls? “Is that even possible?” is the question.

WILLIAM HARTUNG: I think it is, but I think first and foremost for the U.S., we need a defense strategy that’s about defense. If our goal is to dominate the world through technology, these folks are going to win. But if we had an actual defense strategy that said, “We’re not going to fight these unnecessary, illegal, stupid wars, we’re going to educate our population, we’re going to take care of our public health and the military is for extreme circumstances where we’re actually under threat.” And that environment, even if we had this stuff, it would do less damage. But in terms of regulation, I think it’s got to be rigorously tested, which means the government needs people who understand the technology.

I would like these companies to be just another vendor. If you make something that works, we’ll buy it, otherwise shut up. You shouldn’t be running our foreign policy, shouldn’t be manipulating our budget, which means a lot of political reforms in terms of the back and forth of people from industry to government. Campaign finance reform, how these companies dominate government panels that make policies that then benefit them the way they fund think tanks that advocate for them. So there’s sort of that political side.

And then on the technical side, we need people who aren’t making bank off of this, evaluating whether it works, which probably means paying at least some small number of folks more in government to be on that side of the table instead of going to get rich in the tech sector or the venture capital sector. And Europe has tried to do more and I think that is changing now with this, “Oh, they must spend 5 percent of their economy on military hardware,” which is not going to protect them from Russia.

We’re not going to fight World War III. It’s going to be cyber attacks and political interference and a few people slipping across the board. It’s not going to be tanks rolling. They’re buying fighter planes in Europe. They’re going to be delivered eight years from now. Is that going to defend them from Russia? Is Russia going to go after a 30-member alliance when they couldn’t beat the country much smaller than them and what they thought was going to be an easy conflict? So the whole thing is a bit upside-down, but I think it is going to come back to questions of regulation.

And even in my congressional district, there’s a guy, Alex Boris, who had got the first AI regulation bill through a state legislature and Palantir was trying to take him down. The founder put out all this propaganda. Boris had worked for Palantir.

So Palantir said, “You can’t trust this guy because he worked for us and he’s behind ICE.” So it was the most diabolical thing ever. And people work in a Palantir are like, “Why is our founder saying we’re a bunch of murderers?” I mean, it was just, they’ll do anything. And in fact, anybody who stands up for regulation of AI is going to get the same treatment from these dark money operations.

I think one thing that works at our advantage is I think they’re so dishonest, they’re so diabolical, they’re so just unpleasant as human beings that they’re more galvanizing of an opposition than somebody who speaks in bland corporate language and slips you a dollar under the table. But that’s small solace if we don’t stop them. But I think there is a lot more interest in doing that and a lot of younger activists, first of all, they understand the technology better.

Second of all, they have a broader vision. They see the immigration crackdown and the wars overseas, especially the war at home, war abroad as linked and then they see that these same companies are profiting from both. For example, there’s an environmental group in South Texas that’s trying to block SpaceX from creating a larger space launch facility by grabbing some public lands. There’s a group in Ohio trying to block Anduril from building a big plant outside of Columbus. So there’s local resistance, there’s reason the tech sector, the resistance because they’re with what they’re doing to our budget. So all of this has to come together in a way that maybe it hasn’t before. And unfortunately, as you know, having better ideas is not enough in this world. If it was, we live in a much nicer planet, but so it’s about organizing and power and it’s a different kind of power we’re confronting and they’re throwing more at us every minute.

We’re kind of still getting our footing, but I think we will. I think we’ll find ways to fight back.

SCOTT HARRIS: Well, it’s a complicated picture, as you say, a lot of dark and then some beams of light as you said. There is local organizing, even on the front of the data centers, people are sprouting up all over the country opposing those data centers.

WILLIAM HARTUNG: Yeah. They started in Madison. There’s a push in New York because these guys make it sound like these things come sprouted out of their head. But they use energy, they use workers. I’m pretty sure they’re not going to have great conditions. They’re all anti-union. A lot of their other productions done overseas. Are they going to do that for the military stuff? We don’t know. And a lot of these cars are built in China. I think once you start looking closely at who these people are, they’re not these disruptors about to save us from some sort of quagmire that we’re in. They’re basically going to get us all killed if we don’t stop them, but I think we can.
SCOTT HARRIS: Again, a hopeful note to end on. William, thank you so much for spending time with us and refer our listeners to your article at the Nation magazine at thenation.com. “The Pentagon is Going AI First” and do you want to leave our listeners with the web address for the Quincy Institute as well?
WILLIAM HARTUNG: Yeah, it’s quincyinst.org, but if you just Google Quincy Institute or my name, it’ll pop up and we have a lot to say. We just need help.
SCOTT HARRIS: Yeah. Well, we’ll stay in touch. I definitely look forward to talking about this and other issues that you’re working on because it’s a real eye-opener. Appreciate it.
WILLIAM HARTUNG: Well, thank you. Even your list of guests is like, “God, I got to listen more often.” So I appreciate it.
SCOTT HARRIS: Thanks, William. Take care. Goodnight.
WILLIAM HARTUNG: Take care. Bye-bye.
SCOTT HARRIS: That’s William Hartung, senior research fellow at the Quincy Institute for Responsible Statecraft.

Subscribe to our Weekly Summary