Over the weekend, Ukraine launched drone strikes against three data centres owned by Yandex, the Russian technology and artificial intelligence giant often described as “Russia’s Google.” Almost immediately, Ukrainian President Volodymyr Zelenskyy stepped forward with a striking claim: the attacks had silenced a vast army of pro-Russian bots. Speaking by video link to the Warsaw Security Forum on Monday, he said that suddenly, on many social media platforms, it had become easier to breathe. “A lot of people noticed it,” he told the audience. “There was less hate, less division and fewer provocations. Because there were fewer bots.” For many people who had grown exhausted by the endless, toxic flood of comments, the idea was intoxicating: a few well-placed bombs could knock out the machinery of online manipulation and make the internet feel human again. When one of those services disappears, the reasoning went, so too do the puppets that danced to its strings. But as the story spread across forums and feeds, some researchers who study disinformation and digital warfare began to wave red flags. They asked a simple, uncomfortable question: Is it actually true? And even more importantly, does believing it make us safer, or does it simply let us feel better while the real threat continues to operate in the shadows?
The answer, according to experts who study these networks, is a lot messier than the initial headlines suggested. Yandex is indeed a massive player in the Russian digital ecosystem—it runs the country’s most popular search engine, offers taxi hailing, banking, email, cloud storage and mapping, and has poured enormous resources into artificial intelligence development. When its data centres were hit, the effects were tangible. According to the independent Russian news outlet Astra, users in dozens of Russian cities, as well as in Kazakhstan, Belarus and Armenia, found themselves unable to order taxis or access banking services. For a country that depends heavily on Yandex’s integrated services, this was a genuine disruption. But whether that disruption translated into a meaningful blow against Russian propaganda operations is another matter entirely. Lukasz Olejnik, a visiting senior research fellow at the Department of War Studies at King’s College London, immediately took to X to warn against what he called a “dangerous misconception.” The “bots went silent” story, he wrote, is misinformation. It tells people the threat is gone while influence operations keep going. Samuel Woolley, a researcher at the University of Pittsburgh who studies propaganda and disinformation, echoed that caution. In an interview with CBC News, he said he had not seen any research or data that actually tracks all the propaganda bots online, and called it a misconception to suggest that all Russian propaganda bots went down because of the strikes. In other words, the desire to believe in a quick victory over the bots may have created a comforting story that obscures a far more complicated and stubborn reality.
Part of the confusion, both experts explained, comes down to the word “bot” itself. For most of us, the term conjures up images of fake accounts on social media—username and profile photo, spewing outrage or support for the Kremlin under the direction of a troll farm. But in technical terms, a bot is simply any kind of software that completes automated tasks on the internet. Search engines like Yandex—and Google, for that matter—use bots called web crawlers to scan and index the internet. These crawlers are what allow a search engine to find new pages and rank results. They are not posting divisive comments or trying to sway your opinion on Ukrainian statehood. They are just doing the unglamorous, behind-the-scenes work of making the search engine function. After the Ukrainian strikes, people on social media were quick to share a chart from Cloudflare, a U.S. company that monitors global internet traffic. The chart seemed to show Russian bot activity dropping to nearly zero. But Olejnik pointed out that the Cloudflare data merely showed a decline in automated traffic from Yandex’s search engine crawler—not a decline in the kind of bots that spread disinformation by posting inflammatory messages on Twitter, Facebook, Telegram or TikTok. Treating those two things as the same is a category error that, in this case, made for a very misleading narrative. Woolley added that it makes perfect sense that when you physically attack Yandex, automated traffic related to Yandex will drop. The outage affected the company’s servers, so all the little programs that were scraping and indexing websites suddenly had nowhere to go. That’s not the same thing as a network of trolls deciding to take the day off. It’s just the internet equivalent of unplugging a vacuum cleaner and mistaking the sudden quiet for a lack of dust anywhere in the house.
Woolley also reminded people that automated traffic is far more common than most of us realise. Some estimates suggest that over half of all online traffic is automated. But that doesn’t mean that over half of the internet is made up of malicious, manipulative bots. Most of that automation is harmless or even essential—search engines crawling websites, online stores checking inventory, spam filters sorting through email. The word “bot” has become so loaded with negative meaning that people tend to assume all automated activity is somehow sinister. But the reality is that a lot of it is just the plumbing of the internet. What’s more, even the malicious side of Russian information operations is much larger and more flexible than any single company’s server farm. Olejnik noted that Russian influence operations involve thousands of accounts across a wide range of platforms, and that they have proven themselves capable of adapting to disruptions. They hide behind proxies, hijacked devices and even Western cloud services, he wrote on X. They are still operating. Taking down a few data centres, however symbolic or strategically useful in other ways, is not the same as dismantling the machinery of disinformation. The propaganda network is not a single building with a single generator. It is a sprawling, decentralised, endlessly adaptable operation that has weathered sanctions, takedowns and countermeasures for years. Suggesting otherwise, experts argue, not only misreads the data but also creates a dangerous false sense of security at a time when vigilance is still crucial.
Even as the “bots went silent” narrative spread, it seems Zelenskyy’s own team began to soften the claim. His office later told the BBC that the president had not actually been referring to a decline in bot activity. Instead, they said, he was talking about the general atmosphere on social media—that he had seen reports from many users about fewer provocative and hateful comments. This is a meaningful distinction. An atmosphere can change for a lot of reasons: a temporary outage of a major server, a shift in attention due to breaking news, or simply people choosing to log off after a stressful week. It does not necessarily mean that the underlying machinery of disinformation has been dismantled. The clarification suggests that even the Ukrainian leadership understood that the initial interpretation had gone too far. But the story had already captured the public imagination. It was too satisfying, too neat, too much like a victory in a war where victory often feels impossible. The idea that a drone strike could silence the voices of hate online is an appealing fantasy in an era when those voices seem louder than ever. But as Woolley and Olejnik both stressed, the real threat is more diffuse. It lives not only in Yandex’s servers but in thousands of coordinated accounts, in hijacked devices, in influence operations that shift their tactics as soon as they are discovered. The fight against disinformation is not a one-time strike; it is a long, grinding struggle that requires constant attention from researchers, platforms, governments and ordinary users.
In the end, the episode offers a powerful lesson about the way we consume information in the digital age. We are so hungry for good news, so desperate for a sign that the battle against manipulation is being won, that we can easily mistake correlation for causation and a chart for a smoking gun. When the data actually says something more subtle, our own biases can fill in the gaps with a story that feels right. That is not to say the strikes on Yandex were meaningless. They disrupted an important Russian company, sent a message about the reach of Ukrainian drones, and may have temporarily complicated some operations. But resilience is the defining feature of modern disinformation networks. They are built to survive shocks. They have redundant systems, global footprints and a deep understanding of how to exploit human psychology. The threat did not disappear when Yandex went offline, any more than the sea disappears when you close a tap. What matters now is that we learn to read the data more carefully, to question our own assumptions, and to avoid the comfortable myth that a single attack can silence a machine that was designed to keep talking. The bots are still out there, in some form, waiting to be relaunched. The best defence is not a single strike, but a clear-eyed understanding of how influence operations work, a willingness to resist easy narratives, and a collective commitment to making the internet a place where hate does not get to set the agenda. That, far more than any drone attack, is what would make it easier to breathe.