Are The Killer Robots Inevitable?

Summary

What A Day – Episode Summary
Episode Title: Are The Killer Robots Inevitable?
Host: Greg Walters (in for Jane Coaston)
Date: June 5, 2026

---

Episode Overview

This episode dives into the realities and myths around the use of artificial intelligence (AI) in the American military, prompted by President Trump's administration doubling down on AI-driven warfare capabilities. Host Greg Walters interviews Mike Horowitz, a former Deputy Assistant Secretary of Defense and University of Pennsylvania professor, to unravel both the hype and legitimate concerns about autonomous weapons—so-called “killer robots”—and the challenges of regulating this rapidly evolving domain. The conversation sets aside science-fiction fears to deliver nuanced analysis of practical AI applications, existing policies, and the U.S.–China arms race in AI.

---

Key Discussion Points & Insights

---

1. Setting the Stage: AI and the Military [00:04–02:33]

• The episode opens with news—sometimes absurd—before pivoting to the main theme: the Trump administration's push to expand AI use in military settings.
• Greg Walters highlights both the enthusiasm and growing caution inside the Defense Department about AI's use on the battlefield.
• Admiral Frank Bradley, head of US Special Operations: “We need to be, quote, very careful about how AI is employed by the military and that we need to have confidence that it's going to deliver violence only where we intend it to be delivered.” [01:42]

---

2. Debunking the 'Terminator' Myth [02:33–03:31]

• Walters asks if there are “Terminators” or killer robots hidden in the Pentagon.
• Mike Horowitz clarifies: The military is nowhere close to building humanoid killer robots; this anxiety reveals more about human psychology than technical reality.
  • "There's no need for concern. There are no Terminators in the basement of the Pentagon." – Mike Horowitz [02:49]
  • "There’s no Deputy Assistant Secretary of Skynet or something." – Greg Walters [03:17]

---

3. How AI is Really Used in the U.S. Military [03:31–05:13]

• AI’s primary functions are often mundane, including HR, logistics, and synthesizing enormous streams of data.
• While battlefield applications exist, the bulk of AI in defense solves data and collaboration headaches.
  • “So much of it is the boring part… The vast majority of the Pentagon's investment in artificial intelligence is actually to address issues that the Pentagon has... There's almost no organization in the world with more information and more information that is stovepiped...” – Mike Horowitz [04:34]

---

4. Autonomous Weapons: Distinguishing Fact from Fiction [05:13–08:16]

• Walters raises public concern about weapons that can pick out and shoot targets on their own.
• Horowitz explains: Autonomous weapons are not new—examples like the Navy’s Close-In Weapon System have existed since the 1980s, operating on pre-set algorithms, not modern AI.
  • “The U.S. Navy has a system called the Close-In Weapon System... the person operating it can turn on an automatic mode... That's a system the US military has employed since the 80s...” – Mike Horowitz [05:36]
• However, AI can process information and react to hostile environments (like Russian jamming in Ukraine) more powerfully than previous tech.

---

5. Regulating Military AI: The Needle to Thread [08:16–10:36]

• Walters asks: How does the U.S. approach rules for AI in warfare?
• Horowitz underscores a dilemma—overregulation could make America fall behind countries like China, but underregulation could unleash dangerous, uncontrollable systems.
  • “There are risks to overregulation... you don't want the US Military to end up like... rolling its horse cavalry up against, you know, like, Nazi tanks… There's also a lot of risk from underregulation... those kind of go haywire on the battlefield in a way that causes accidents and miscalculation.” – Mike Horowitz [08:25–09:36]

---

6. Trust, Policy, and the Human Element [09:36–12:51]

• Walters probes how Americans can trust administration policy and oversight.
• Horowitz references DoD Directive 3000.09, which sets guardrails for autonomy in weapons, aiming to ensure systems are tested, safe, and human-responsible.
  • “There should always be a human responsible for the use of force. No matter what. Period.” – Mike Horowitz [06:51]
• If leadership ignores the rules, soldiers themselves are likely to resist unsafe systems since their lives are at risk.
  • “Unsafe AI systems don’t work. And generally nobody fights harder against those than the military.” – Mike Horowitz [10:31]
• The conversation concludes that integrating AI is inevitable—what matters is adapting regulation as technology evolves and maintaining deliberate, thoughtful oversight.

---

Notable Quotes & Memorable Moments

• “Does the Pentagon have a Terminator yet?” – Greg Walters [02:35]
• “Every time in the movies you give a robot a brain and a gun, it seeks to destroy humanity... but there's no Terminators in the basement.” – Mike Horowitz [02:49]
• “Almost no organization in the world with more information... that is stovepiped... More collaboration needed.” – Mike Horowitz [04:34]
• “The critical thing… always there should be a human responsible for the use of force.” [06:51]
• “We don't want the US military to end up like… rolling its horse cavalry up against, you know, like, Nazi tanks.” – Mike Horowitz [09:12]
• “We especially need a framework for guardrails surrounding artificial intelligence that adapts as the technology changes.” – Mike Horowitz [12:18]

---

Timestamps for Key Segments

• [00:04] – Episode start, intro to killer robots topic
• [01:42] – Admiral Frank Bradley’s caution on AI
• [02:33] – Greg introduces Mike Horowitz; Terminator myth dispelled
• [03:31] – Real AI uses in the Pentagon
• [05:13] – Autonomous weapons: US history and tech
• [08:16] – The dilemma of regulating military AI
• [09:36] – Trust, policy, and who ensures rules are followed
• [12:51] – Final thoughts on AI inevitability and the need for adaptive regulation

---

Listener Takeaways

• “Killer robots” are more science fiction than present reality, but autonomy in weapons is not new.
• Much of military AI investment targets mundane but critical data and decision-making bottlenecks.
• True “human-in-the-loop” policy—requiring human responsibility—is official Pentagon doctrine, but practical and ethical challenges remain.
• The U.S. must balance speedy AI adoption (to compete with China) and the risks of unsafe, unregulated technology.
• Thoughtful, adaptive guardrails and honest oversight are crucial as AI’s capabilities grow and redefine modern warfare.

---

Tone & Language

The conversation balances humor and gravitas, using pop culture jokes and real-world examples to make a complex, often scary topic approachable and clear.

---

Episode in a Sentence

A nuanced, un-hyped breakdown of what “killer robots” really mean for the U.S. military, why the debate is complex, and why getting the rules right is more urgent than ever.

Transcript

[Greg Walters]

Happy Friday, June 5th. I'm Greg Walters in for Jane Coston, and this is what a day. The show that's gotta hand it to Colorado GOP gubernatorial candidate Victor Marks for batting away questions about his extraordinary military exploits by pointing to his dog on stage and saying, basically, if you don't believe me, ask my dog. This little dog. She's gonna go bite you right now. Kyle, she was in Syria and Iraq. So is she lying, too? Well, the dog's not running for governor. You're running for governor. I don't know, man. I think the dog's got my vote. On today's show, former Attorney General Pam Bondi throws Acting Attorney General Todd Blanche under the bus. And are UFOs really demonstration? One exorcist seems to think so. But let's start with killer robots. Or, to be more precise, the Trump administration's push to supercharge the American military with artificial intelligence. AI isn't just clogging up your Instagram feed with cartoon cat romance, revenge videos and helping your cousin pass AP European history class. It's also on the battlefield. The US Military is leaning into this powerful new technology from Venezuela to Iran. Back in December, Secretary of Defense Pete Hegseth put it this the future of American warfare is here, and it's spelled AI. As technologies advance, so do our adversaries. But here at the War Department, we are not sitting idly by. Under the leadership of President Trump, America

[Mike Horowitz]

will lead the charge on this technological

[Greg Walters]

transformation by revolutionizing the the way we win. But not everybody is so gung ho. Some are arguing we need guardrails in place. We've even heard notes of caution from those in uniform. Admiral Frank Bradley, head of US Special Operations Command, recently said that we need to be, quote, very careful about how AI is employed by the military and that we need to have confidence that it's going to deliver violence only where we intend it to be delivered. And, yeah, I'd agree. To learn more, I spoke with someone who worked in the Pentagon on military AI systems under former President Joe Biden. Mike Horowitz. He's a former Deputy Assistant Secretary of Defense and a professor at the University of Pennsylvania. Mike, welcome to what a Day.

[Mike Horowitz]

Thanks so much for having me.

[Greg Walters]

So you were deeply involved in the Pentagon's AI program back in the Biden days. And so let me start by asking this. Does the Pentagon have a Terminator yet?

[Mike Horowitz]

I mean, it depends on what you mean by a terminator, but really the answer is is no, and not even close. I think it says more about how we think about ourselves that every time in the movies you give a robot a. A brain and a gun, it seeks to destroy humanity than it does about the robots themselves. But there's no need for concern. There are no Terminators in the basement of the package.

[Greg Walters]

There's no deputy Assistant Secretary of Skynet or something.

[Mike Horowitz]

I mean, what a wonder that would be, except for the fact that Skynet would destroy humanity regardless of what political party you're from. So very little upside for anybody in that scenario.

[Greg Walters]

So how is artificial intelligence being used by the US Military right now?

[Mike Horowitz]

The way I would think about artificial intelligence in the US Military is to understand that it's a general purpose technology. And everywhere from the back office to the battlefield, there are uses of artificial intelligence. So that includes all of the uses of AI that any company around the world might use it for. Hr, logistics, like varieties of paperwork kind of things, as well as things like processing intelligence information, trying to separate the signal from the noise, understand what's happening with all of the disparate sources of information the Pentagon gets, and then up into including the sort of pointy edge of the spear and the battlefield, the potential uses of artificial intelligence both to speed up what the American military does and to make it more accurate and effective.

[Greg Walters]

So everything from the very boring stuff that maybe we're all used to with AI to things that are like synthesizing information, choosing targets, thinking faster than the other side.

[Mike Horowitz]

So much of it is the boring part. And that's one of the things I think is difficult to communicate at times. Like, the vast majority of the Pentagon's investment in artificial intelligence is actually to address issues that the Pentagon has. There's almost no organization in the world with more information and more information that is stovepiped in a way that makes it almost impossible for everyone that needs to access it, to be able to access it. So most of what AI does for the Pentagon in some ways, and frankly will do for a number of years, is trying to solve a bunch of those data and information issues that have made it really difficult to collaborate, even though sometimes they're good for secrecy.

[Greg Walters]

But yet there are things that people are concerned about. And I think one of the key issues that I've heard people talk about with regard to AI and warfare is the question of autonomous weapons and whether we may be going in that direction, how quickly we get there. An autonomous weapon will be able to pick out its own target and decide whether or not to shoot. Now, is that something we should be worried about?

[Mike Horowitz]

So an autonomous weapon system is Something the American military has actually been fielding since the early 1980s. And that complicates this question enormously. Let me break that down. The U.S. navy has a system called the Close in weapon System. It's like an enormous Gatling gun, like from like post Civil War kind of era that has a little algorithm programmed into it so that if there are all these threats coming to hit, like one of our aircraft carriers or other ships, the person operating it can turn on an automatic mode and by algorithm it can then select and engage targets. And that's a system the US military has employed since the 80s that protected the Navy in the Red Sea and during the houthi strikes that 40 different countries around the world have. Now that's not obviously a large language model, that's not like ChatGPT or Claude or something like that. But one of the things that makes this conversation so difficult is how much autonomy is built into modern weapon systems already in ways that people have been very supportive of because they've decreased collateral damage, protected American soldiers, et cetera.

[Greg Walters]

I mean, fair, but AI is different, right? I mean, it's potentially orders of magnitude more powerful than that kind of technology, isn't it?

[Mike Horowitz]

Yeah, absolutely. And that's why I would draw a distinction in some ways between the decision to use force and then the way that that decision is implemented. The critical thing, from my perspective at least, and this is also what Pentagon policy says, like hopefully they're following it, is that always there should be a human responsible for the use of force. Like no matter what, period.it frankly doesn't matter whether it's like a spear or a sword or a radar guided missile or some fancy AI enabled system, there should always be a human responsible for the use of force. That's different than say, what is the guidance system on board a missile. And there, there are really technical operational reasons for increasing autonomy on the battlefield. Like for example, all the, all the drones that Ukraine fires at Russia to try to, as they try to defend themselves from Russia's continuing invasion. A lot of those just drop out of the sky because of Russian jamming, what experts often call electronic warfare. What those that jamming is trying to do is disrupt the data link between the Ukrainian operator of that attack drone and the drone itself. But if you don't need the data link because you already, you have a, you know, a sort of like more powerful than it used to be, say like algorithm on it that can still go and hit the target. That's the kind of thing that Ukraine is doing to protect itself from Russian attacks right now.

[Greg Walters]

So how do we think about, and I mean we, the United States, think about making rules to govern this new frontier in warfare.

[Mike Horowitz]

It's so challenging because in some ways, there are risks to overregulation of artificial intelligence from a military perspective as well as underregulation.

[Greg Walters]

Break that down for me.

[Mike Horowitz]

Yeah, Let me take each of those briefly in turn. From an overregulation perspective, the risk is that the US Military falls behind competitors like China, in Russia, like, really especially China when it comes to adopting artificial intelligence, because they're like full speed ahead, nonstop. And there are lots of advantages to AI integration in terms of processing information faster, being more accurate, being more effective. So there are real risks in falling behind. You don't want the US Military to end up like, like an army in World War II rolling its horse cavalry up against, you know, like, Nazi tanks.

[Greg Walters]

It doesn't sound great.

[Mike Horowitz]

There's also a lot of risk from underregulation, though. It's that you deploy a bunch of autonomous weapon systems or AI enabled systems that aren't safe, that you don't have full control over, and those kind of go haywire on the battlefield in a way that causes accidents and miscalculation. And so there are dangers in both directions, which is one reason why there's a really difficult needle to thread here.

[Greg Walters]

A moment ago, you said that it was important that this administration is following its own policy. What do you say to people who might be concerned that they shouldn't trust this administration to do what it says it's going to do?

[Mike Horowitz]

It's a really tough question, you know, to be honest. You know, when I was in the Pentagon, one of the things that my office did, or the office I was honored to run was write what was then called Department of Defense Directive 3000.09, Autonomy and Weapon Systems, which sets very clear goes guardrails and rules surrounding the Pentagon's development of autonomous weapon systems, essentially to ensure that if the Pentagon is to develop autonomous weapon systems that they work, which is in the interest of the military, since, like, nobody wants their technologies to work more than our soldiers because it's their lives on the line when these systems fail. The challenge is that when you, if you have an administration that isn't following the rules in general or what ensures then that those are followed in this, in these cases, I don't have a great, like thing to tell you in this context, to be perfectly honest, except for the incentives of our soldiers, like I was saying, and that the good thing in the AI case is that unsafe AI systems don't work. And generally nobody fights harder against those than the military. And so if you had some kind of push to try to field a bunch of unsafe like military AI systems, I suspect you'd get, at least I would hope you'd actually get really strong pushback from our soldiers because they wouldn't want to use those systems. Maybe that's wishful thinking, but I think it's important, it's important to keep that in mind.

[Greg Walters]

Well, I think a lot of people feel that there are some things they hope this, this administration will be doing that it may or may not be doing. But it sounds like you're saying that in the end it's really important that we get this right, regulation, right, that we think hard about it and do this deliberately.

[Mike Horowitz]

We have to be deliberate when it comes to thinking about the integration of artificial intelligence. I think for two reasons. One, this is a general purpose technology. We look at our phones, look at society, look at the economy. The military is going to be no different. And so artificial intelligence integration into something like the American military is inevitable. The question is essentially how we do it. The second thing is the US is in what I would call essentially a generational competition with China for leadership across all dimensions. One of the main dimensions of that competition is in artificial intelligence. And given that there are risks to overregulation and under regulation, we need to be really thoughtful about it. And we especially need a framework for guardrails surrounding artificial intelligence that adapts as the technology changes. Because as AI technology improves, you know, like as you, as you trust whatever LLM out there, like you might use more because it actually works. We should use it for more things. But also you don't want to use it if it's not going to work and like hallucinate and make things up. We need to be able to tell the difference and adapt accordingly.

[Greg Walters]

Mike, very thoughtful. Really appreciate you having you here.

[Mike Horowitz]

Thanks so much for having me.

[Greg Walters]

That was my conversation with Mike Horowitz, a former Deputy Assistant Secretary of Defense, focused on the future of war. And here at Whataday, there are no Terminators in our basement. We have real live hosts like Jane, who'll be back on your feed at a new time starting this Monday in the afternoon. So if you like the show and don't want to miss our new PM drop time, make sure to subscribe. Leave a five star review on Spotify and Apple Podcasts, watch us on YouTube and and share with your friends. More to come after some ads.

[Erin Ryan]

What a Day is brought to you by Quint. Summer always makes me rethink what I'm reaching for every day. Lighter fabrics, better materials, pieces that just feel good the moment you put them on and look effortless. That's why I keep coming back to Quince. They focus on high quality essentials. Think breathable linen, soft organic cotton, washable silk, but without the luxury markup. It's that rare balance where everything feels elevated but still easy. Quint has beautiful everyday pieces like 100% European linen pants, dress and tops with styles starting at just $32. Their denim is soft and easy to wear and the organic cotton sweaters are perfect for layering on cool summer nights. Everything at quince is priced 50 to 80% less than similar brands. Quince works directly with ethical factories and cuts out the middlemen so you're paying for quality, not brand markup. And it's not just clothing. Quint has really become a destination for elevated essentials across home, kitchen, bedding and beyond, making it easy to bring a more premium feel into everyday life. I love Quince because I have this hang up where if I have clothes that are nice that I paid a lot of money for, I don't wear them. They just live in my drawer cause I'm afraid to wear them because I'm afraid to get them dirty. But quints looks expensive but if I get it dirty it's easy for me to clean and it just holds up. I love that I can wear quints without feeling like I am endangering my expensive clothes. What I am doing is wearing clothes that last, that feel high quality and that look great. Elevate your summer wardrobe. Go to quince.com wad for free shipping on your order and 365 day returns now available in Canada too. That's Q-U-I-N-C-E.com wad for free shipping and 365 day returns. Quince.com wad this podcast is brought to

[Podcast Announcer/Producer]

you by Planned Parenthood Federation of America. Everyone deserves access to high quality, affordable healthcare. That's why millions of people rely on Planned Parenthood health centers for services like cancer screenings, wellness exams, birth control, STI tests and more. In fact, one in three women in the US have been to Planned Parenthood for care. So it kind of blows my mind that lawmakers are trying to permanently shut them down. They already passed a law that blocks patients that use Medicaid from getting the affordable preventative care they need, and now they want to make that law permanent. Nobody benefits when people are getting fewer exams, paying more for care or just skipping the health care they need. And we know that's happening. Planned Parenthood health centers have reported that fewer patients are getting basic care like breast exams and STI tests since the Trump administration and their backers defunded Planned Parenthood. Patients shouldn't suffer because of their loss. Our communities deserve better to get involved. Text update to 22422.

[Greg Walters]

Here's what else we're following today. Head of Line

[Mike Horowitz]

Cole's a great business, really a big business. And it's real power. In terms of power, there's really nothing like it. You have so many different alternatives. They talk about some, but there's no real alternative.

[Greg Walters]

Ah yes, no real alternative to the great business of coal. It's such a really big great business that the Trump administration has to spend hundreds of millions propping it up. President Donald Trump announced a plan to spend nearly $700 million supporting coal fired power plants and coal exports. During an Oval Office ceremony on Thursday, Trump said the administration will use authority under a Cold War era national defense law to support 13 coal power plants across the country and help build coal plants in Alaska and West Virginia, the first new US coal plant since 2013. Because real men don't just heat their homes, they heat up the atmosphere while they're at it. And only suckers rely on windmills. Most of the administration's prosecutions of Trump's critics and enemies have absolutely flopped, but prosecutors may be about to notch a win. John Bolton, Trump's estranged former national security adviser, has reportedly agreed to plead guilty to a single count of retaining classified information in a deal with the Justice Department. That's according to unnamed sources who spoke to multiple outlets, including the Associated Press. This is a lot less than what prosecutors were originally going for. Bolton was charged with 18 counts of either retaining or disseminating classified information in a case relating to the book he wrote about his time working for Trump, which I read and let's face it, was boring under this deal. A Source told the AP that Bolton would face a fine of $2.25 million and that any prison sentence would be capped at five years and he could avoid time behind bars completely. But that will be up to the judge. Former Attorney General Pam Bondi threw her successor and ex deputy Todd Blanche under the Epstein bus during an appearance in Congress on Thursday. According to a transcript of her remarks, Bondi said, quote, he was in charge of the process and the entire release of the Epstein files. And this is rough timing because Trump wants Blanche, his own former lawyer to take the AG job permanently. Todd Blanche seems to really want this job, but he may want to ask Michael Cohen how handling Trump's legal business tends to work out. If you're the kind of person who loves spending your summer vacation at home because your president's erratic foreign policy has made it too expensive and complicated to go someplace else, well, I've got great news. American Airlines is temporarily suspending some of its routes this summer. Why? Because Trump's war in Iran has sent the cost of jet fuel soaring. Americans said in a statement that it adjusted service for select routes in August and September and that impacted travelers would be offered alternative arrangements or refunds. Now, if you live in LA and want to visit friends in Cleveland, Columbus, Pittsburgh or Washington, D.C. this summer, better make sure your plans are still good. On the other hand, if you didn't really want to go in the first place, maybe the real summer vacay this year will be pizza pieces, PJs and season three of House of the Dragon. Speaking of things you might find in a TV show that you really shouldn't watch right before bed, we'll close with a little news item involving demons and UFOs, which, trust me, as somebody who works in news and podcasts, are two things that don't normally pop up in the same headline every day. The Catholic Archbishop of Washington, D.C. removed a well known priest, Monsignor Stephen Rossetti, from his position position as exorcist of the Archdiocese on Wednesday. Rossetti's removal came after he recently made some interesting comments on a video that was posted to his Facebook page. There's no question in my mind, personally, again, this is not de fide, but it's my personal belief that probably many, if not most of these UFO sightings are in fact demons. But that didn't sit right with the Archbishop, Cardinal Robert McElroy. McElroy said Rossetti's statements connecting UFOs with demons, quote, gravely undermined the Church's very precise teaching on the devil, Demons and exorcism, which I take it is that those things simply do not travel on flying saucers. And that's the news. Before we go on Monday. What a day. Episodes are going to start hitting a little different. They'll be dropping on YouTube and your podcast feeds in the afternoon, but you can listen whenever your heart desires. If you like this show. You know the news never stops. So this new schedule will bring you the freshest updates just in time for your evening commute decompression walk. Or as some of us like to call it cocktail hour. Make sure you're subscribed to what a Day on YouTube, Apple Podcasts, Spotify, and follow Rooked Media on social so you never miss an episode. That's all for today. If you like the show, make sure you subscribe, leave a review, don't give your swimming pool any performance enhancing drugs and tell your friends to listen. And if you are into reading and not just about Trump saying that his paint job on the Lincoln Memorial Reflecting Pool makes it, quote, a swimming pool on steroids like me. What a Day is also a nightly newsletter. Check it out and [e-mail] subscribe I'm Greg Walters and I thought the steroids were for the UFC fight, not the pool.

[Podcast Announcer/Producer]

What a Day is a production of Crooked Media. Our show is produced by Caitlin Plummer, Emily Foer, Erica Morrison and Adrienne Hill. Our team includes Haley Jones, Greg Walters, Matt Berg, Joseph Dutra, Johanna Case and Desmond Taylor. Our music is by Kyle Murdoch and Jordan K. We had help today from the Associated Press. Our production staff is proudly unionized with the Writers Guild of America East.

[Mike Horowitz]

Quick question. Are you politically engaged and spiritually exhausted

[Erin Ryan]

if you said yes to both? Welcome home. I'm Erin Ryan.

[Mike Horowitz]

And I'm Alyssa Mastromonaco, and we're the

[Erin Ryan]

hosts of Hysteria, the podcast for women who care about democracy, culture and not losing their minds in the process.

[Mike Horowitz]

We break down the news, call out the nonsense, and spotlight the women actually fighting back on Capitol Hill, in classrooms and everywhere. The stakes are high.

[Erin Ryan]

It's sharp, honest analysis featuring women's voices with humor and zero hand holding.

[Mike Horowitz]

Listen to Hysteria wherever you get your podcasts and watch full episodes on YouTube.