The man fighting Russian disinformation with data, irony, and viral memes. Interview with Marijn Markus

Information warfare has become a critical front in modern defence tech — and Ukraine is already home to a growing number of promising AI-driven startups in the field, including Osavul, Mantis Analytics, and others. Defender begins its deep dive into this domain with insights from Marijn Markus, a Dutch expert who has been closely following the war in Ukraine and understands the mechanics of information warfare. Bojan Stojkovski sat down with him to discuss Russian propaganda tactics, the lessons Ukraine’s frontline can teach the West, and the evolving role of AI on the battlefield.

A data scientist by trade, a meme lord by the LinkedIn algorithm’s will, this is how Marijn Markus describes himself. If you’re on LinkedIn and closely watched updates on the war in Ukraine, you’ve probably seen Marijn’s posts—sharp, data-driven breakdowns of Russian propaganda campaigns, laced with irony and memes.

The 35-year-old Dutchman has been one of the most ardent supporters of Ukraine and has used his skills as a data scientist, communicator, and online organiser to support grassroots aid efforts, and keep international attention focused on the war.

For Marijn, the Russian aggression on Ukraine is personal, as it affected his close friends, such as Sergei—a gamer from Mariupol he met online, who vanished after the full-scale invasion began. One day they were chatting on Discord about anime and video games, and the next, Sergei stopped replying. That sudden silence stayed with Marijn. It turned abstract geopolitics into a visceral loss. “He’s part of the reason I care so much,” he says. “Because I knew him—and then he was gone.”

A couple of weeks ago, I had the chance to talk with him about his mission while we were on the train from the Polish city of Przemyśl to Kyiv, where we were both headed to attend the Invest in Bravery conference.

Defender Media: When did you first become interested in countering Russian disinformation, and what sparked that shift for you?

MM: I’m a data scientist by trade, but I started out as a social scientist—sociology, criminology, statistics. It wasn’t sexy back then. Now we call it “AI” and “data science,” and suddenly, I’m sexy—flips hair. I also did a minor in media and communication—aka propaganda science. I studied the Frankfurt School, who analysed Soviet propaganda. The “firehose of falsehood” tactic, for example—not to convince you of a lie, but to drown the truth in noise.

I remember reading that for class, looking at the TV in 2014, and MH17 had just been shot down. And Dutch news said, “The Ukrainians say the Russians did it. The Russians say the Ukrainians did it.” Also NATO, the Jews, the Illuminati, and so on. And I looked at that book—and it was the same thing. Like studying dinosaurs and seeing one walk past the window.

That changed my life. We call it “fake news” now, but the tactic hasn’t changed in 80+ years. It’s still accusing the other of what you’re guilty of. Call them Nazis. Call them gay. We’re all gay Nazis according to Russian propaganda. It wasn’t obsession—it was relevance. The thing I studied because I found it interesting turned out to be incredibly real.

That was during the big data revolution, when we realised how much information we had access to. That’s how I got into data science. I spent weeks on Twitter, posting nonstop about MH17, verifying footage, analysing Russian comms. People kept saying, “We’ll never know,” which is the most classic disinfo tactic. Just pump out enough lies until people give up. That’s not about believing the lie—it’s about not believing anything at all. That’s what “post-truth” really is: the erosion of meaning. Once you stop believing in truth, you stop caring. If you believe you can’t know, then you do nothing. That’s what they want—paralysis.

This tactic goes back hundreds, if not thousands of years. And that’s why I care so deeply—not just about disinformation, but about information. It’s not about being 100% right. That’s impossible. As a statistician, the goal is never to be right—it’s to be less wrong. That’s the whole point of science. But the moment you shrug and say, “We’ll never know,” you abandon all progress.

People say “but we need to understand both sides,” and it’s like—no. This is a textbook. The Nazis and Soviets invaded Poland the same way—claiming to protect local populations. Same with Finland. False flags, bombing your own people, blaming the enemy. All of it, 80+ years ago. It keeps happening because we’re terrible at history. And that’s why I write about it—so we remember.

In a time when we have the tools to verify almost everything, why do you think so many people still fall for disinformation—or choose not to check the facts at all?

It’s all still very new to us. ChatGPT has only been around for a few years. Google has been here for maybe 20 years. Most people are still technologically illiterate, and yet we have all these tools at our disposal to look things up in ways we couldn’t decades ago. The internet has completely changed the way we live, but we’re still immature in how we use it—and it’s going to take decades for us to mature.

That’s part of why we’re failing so hard. When the steam engine arrived, you think there weren’t a lot of accidents? When radio waves came and we didn’t know what radiation was, you think there weren’t accidents? When electricity arrived, how many people stuck their fingers in sockets and got electrocuted?

“Tech is never the problem. Legacy is. Legacy thinking. Legacy processes. Legacy data. Legacy humans.”

This is the next technological revolution—after steam, after radio, after electricity—it’s the internet and open access to information. Twenty years ago, we called it the digital revolution because companies started digitising their processes. Ten years ago, we spoke of the data revolution because people began analysing massive amounts of data. Today, we talk about the AI revolution—because now we can let models do a lot of that analysis for us.

But the truth is, most companies and organisations never did their homework. They’re still stuck in the digital revolution—they never properly digitised their processes, so they don’t have the data. That means they missed the data revolution too. And now, with the AI revolution, everyone’s screaming “We want to build our own GPT model on our own data!” But do you have your own data? “Yeah, but it’s shit.”

So because people didn’t do the work decades ago, they’re now stuck. We should not confuse cause and effect. A lot of the issues we see with tech today come from that failure to adapt—not because the tech itself is bad. I reject that. I’m a nerd, a techie. But tech is never the problem. Legacy is. Legacy thinking. Legacy processes. Legacy data. Legacy humans.

What will it take for Western militaries to catch up with Ukraine—not just in tech, but in mindset?

I speak to the Dutch military—we’re five to ten years behind what’s happening in Ukraine. And yeah, sure, we have a five-year innovation plan, but we don’t have five years. Right now we’re trying to rebuild our militaries, because all NATO countries need to rebuild theirs by buying back outdated stuff. Oh, we sold off our hand-carried artillery, whatever, blah blah, and now we’re going to rebuild it? Sorry, but that artillery is already outdated, and you’re going to spend millions of our tax euros on rebuilding something that’s already obsolete.

When I speak to the Ukrainian military, they get tech. Two-thirds of that force isn’t professional soldiers—it’s developers, engineers, and software guys who went into the military to defend their damn country. When I speak to the Dutch military? Half of them have never touched a computer with a stick.

There’s a huge tech gap. In most companies, one-third of the staff work in IT. One-third of all job openings are in IT. My mom still says I “do something with computers” like it’s niche, but one in three people here do something with IT. And yet in the military? We don’t have a third who understand tech. Not even close. That’s going to cost us. We’re preparing for tomorrow’s war with people who only know the wars of the past—Cold War doctrine, asymmetric warfare. We no longer know what it’s like to fight another full-on military. We only know how to bomb some militants in Afghanistan. Pardon my French, but that’s the benchmark.

“The front line is a 15-kilometer-wide death zone where anything visible gets hit by an FPV drone. We, westerners, are not ready for that war.”

Just like we once had to shift to asymmetric warfare, now we need to shift back. Ukraine is the only country in Europe today with real, live experience in full-scale military-on-military warfare. And yet we’re still out here talking about buying hand-carried artillery with some XYZ-kilometer range. It’s irrelevant. The front line is a 15-kilometer-wide death zone where anything visible gets hit by an FPV drone. We are not ready for that war. Not with our outdated tactics. Not with our outdated standards.

The summer 2023 Ukrainian offensive didn’t fail because Ukrainians lacked the will—it failed because we told them to use NATO tactics without giving them NATO air superiority. But NATO doctrine depends on air superiority. So we told them: “Do the thing—without the thing.” Charge with your Leopards? No. That doesn’t work. And even now, two years later—it still doesn’t work. All the tanks? Gone. All the artillery? Gone. And most of the troops? Trained on that outdated stuff. It’s irrelevant.

And here’s the crazy part: not only do all NATO militaries need to rebuild their weapons, munitions, and armor, they’re also working with outdated gear and outdated people. Outdated gear you can replace—some people will even make a lot of money off that. But outdated people? That’s a whole different ballgame. You need experienced people to train the next wave. And guess what? There’s only one country we can really learn from: Ukraine.

Denmark gets it—they’re sending troops to Ukraine to learn about drones. Everyone should be doing that. Who else are we going to learn from? Russia? China? China has a strong proxy in Russia. They’ll exploit all that drone experience. But we could match it—or even surpass it—on our side. So instead of questioning whether Ukraine should join NATO, we should be begging them to join. Ukraine is a military asset in ways no other country is right now.

But instead, I find myself stuck in conversations where people say, “We need to build a drone battalion.” No. You need drones in every battalion. One of the tank riders should be flying a drone, looking from above. Artillery should be coordinated by drones. Medivacs should be done by drones. Every battalion needs drones. If you say, “Let’s start with a drone battalion,” then you just don’t understand how serious the situation is. Or your politicians aren’t giving you the means to get serious—which is another issue entirely.

“Instead of questioning whether Ukraine should join NATO, we should be begging them to join. Ukraine is a military asset in ways no other country is right now.”

And yeah, I’ve worked with major Western organisations who say they don’t want to develop killer drones. They don’t want their hardware or software used to dehumanise or kill. They call it “killer AI.” I work a lot with AI companies—they don’t want their visual recognition tech to be used like that.

But look—twelve years ago, Andrew Ng at Google trained a model to recognise cats and dogs. Today, that same tech is in every self-driving car and in every autonomous military drone. We taught a computer to see in 2012. Now, in 2024, it’s a weapon. That’s how fast things go when the pressure is real.

But instead of adapting, we say “we need to regulate.” Sure. But the same companies that say they won’t touch this tech today? The moment the first French soldier dies, they’ll drop that policy. By then, it’ll be too late. That is the real issue. We’re clipping our own wings now. And when the real shit hits the fan—when the Baltics get invaded—it’s going to be too late. 

This is short-term moral comfort at the cost of long-term existential risk. And we’ve been under-investing in our defence for far too long.

Given Ukraine’s success against Russia’s Black Sea Fleet without having a navy of its own, what does this reveal about the future of naval warfare—and why is it the most unpredictable and dangerous frontier today?

In Ukraine, for the first time, we’ve seen modern land warfare between two armies in ways we haven’t before. It’s basically: forget your tanks, forget your armor—it’s a 15-kilometer kill zone made of FPV drones. We’ve also seen modern aerial warfare, which is basically: nothing flies because it gets shot down. There is no aerial dominance—only mutual denial. But the one thing we haven’t seen is modern naval warfare.

The only country with actual experience in modern naval warfare is Russia—and we just saw them lose a third of their Black Sea Fleet, blown up by tiny boat drones that Ukraine assembled. Now, the rest of their ships are hiding in port, too afraid to venture into the Black Sea. Ukraine, meanwhile, has no navy of its own, but it still kicked the Russian Navy’s ass.

The big unknown is what Navy-on-Navy warfare looks like in modern times—because Ukraine doesn’t have any cruise ships or equivalent large vessels. That’s what makes future naval warfare the scariest right now: we have the least data on it.

What we have seen is that combined arms warfare has become a huge differentiator—NATO’s doctrine of integrating land, air, and sea. Ukraine is doing this already. We see naval drones going all the way to Crimea and launching aerial FPV drones from there to strike deep behind enemy lines. The boat, which has satellite connection, functions as a transmitter for the remote-controlled drones. Anyone could just sail into a port and launch FPV drones that start blowing up high-value targets. This isn’t sci-fi—it’s already happening.

“The next big leap is AI-controlled naval drones—underwater ones.”

Now imagine doing this with aerial drones as delivery systems. You could have winged drones fly into a city and drop FPV drones to take out multiple targets. They could do that to The Hague. And we’re not ready. Sure, if we see them coming, we could blow them up—with expensive missiles. But you can’t scale that. You can’t shoot down $1,000 drones with $1 million missiles indefinitely. In a real war, especially against China, you’ll lose. The cost to defend exceeds the cost to attack. They’ll just keep coming. You can’t jam it either, unless you jam your entire airspace—which is also unworkable.

So we shouldn’t just be asking what naval warfare will look like—we should be looking at how land and aerial warfare are merging. And the next frontier is underwater. Why haven’t we seen underwater drones yet? Because water blocks radio waves. You need satellite uplink, which works fine for crewed submarines that surface to communicate, but not for autonomous drones. Unless they’re fully AI-controlled.

And we’ve already seen AI-controlled aerial drones and some land drones. The next big leap is AI-controlled naval drones—underwater ones. Why stay vulnerable on the surface if you can go underwater and become a torpedo-like AI drone?

Bojan Stojkovski

Bojan Stojkovski is a freelance journalist. Based in Skopje, North Macedonia, he reports on technology, science, and environmental issues, as well as post-war societies in the Western Balkan countries. His work has been featured in Foreign Policy, WSJ, ZDNET, New Eastern Europe, and Interesting Engineering.

• X
• LinkedIn