Drone Strikes Didn’t Take out a Russian Bot Farm, But We Were Primed to Believe It

On Monday, Ukrainian president Volodymyr Zelenskyy claimed that a Ukrainian bombing strike on Yandex’s data centers, a large Russian technology company, had led to a sudden decrease in bot activity online. “When our strikes reached Yandex’s data centers, it suddenly became a little easier to breathe on many social media platforms,” Zelenskyy declared, speaking at the opening of the Warsaw Security Forum. There was, he added, “less hate, less division, and fewer provocations because there were fewer bots. And it’s pretty clear where those bots were coming from.” 

The claim quickly picked up speed, and gained specificity: that following the Ukrainian strikes, all kinds of internet users were suddenly less present online, including MAGA accounts, those supporting Reform UK, and ones supporting AfD, the German ultra-right party. A Polish political scientist shared claims that bot activity supporting AfD had dropped by 56 percent, a post that was viewed millions of times and has since been deleted. As Radio Free Europe has reported, a Czech politician shared a similar claim, writing, “So much talk about German patriotism…then the connection from Russia goes down.” 

The underlying claim, in other words, was that a great deal of political activity online across a number of countries is inauthentic and led by Russian bots. Their sudden alleged silencing was a simple, feel-good story amidst the endless fight against online disinformation.

The only problem was that it simply wasn’t true.

Tech experts quickly explained that the “bots” that had been taken offline by the Yandex strike were actually bot crawlers, also known as web crawlers, which continually browse the internet in order to index web pages. A widely circulated chart from Cloudflare, the infrastructure company, did indeed show a drop in traffic from Yandex’s servers, which made sense; both the company’s web crawlers and search index requests were disrupted by the bombing campaign. 

The first issue, then, was the word “bot,” which “means different things in different contexts, which is contributing to some public confusion,” explains Renee DiResta. She’s an associate research professor at Georgetown University’s McCourt School of Public Policy, a social media researcher, and a globally recognized expert on disinformation. “It is used as shorthand for both web crawlers and online trolls. Web crawlers plummeting makes sense, but online trolls usually have more resilient infrastructure.” 

The second problem, however, was that the ultra-specific claims about less activity in certain areas—MAGA, pro-Kremlin bot accounts, and AfD—had absolutely no basis in fact. 

“We see no “dramatic drop” in traffic of Kremlin-aligned human trolls-for-hire targeting Russian domestic users (which we have been tracking for years),” wrote Antibot4Navalny, a group of anonymous researchers who track and respond to Russian disinformation. 

“I’m sorry to burst everyone’s bubble, but as of now, there is no solid evidence that Russian bot farms have become inactive due to the Yandex data centre strikes,” agreed Vatnik Soup, another group that monitors Russian government-backed disinformation. “This doesn’t mean the strikes haven’t had an effect, but no one has demonstrated it with solid data so far,” the account added. “Many users have reported seeing fewer pro-Kremlin posts and replies since the strikes, and some automated pro-Russian and pro-MAGA accounts have reportedly gone silent. But this could simply be a coincidence.” 

“We don’t see it,” tweeted Valedmar Osted, who works at the Scandinavian tech and social media advisory firm Common Consultancy. “We took 76 AfD politicians and 136 MAGA accounts on Facebook. We compared each post with that account’s normal. Same post age. Same weekday. Same time of day. AfD after the strikes: −8% to +5%. MAGA: +4% to +10%. That’s ordinary day-to-day noise.” 

Antibot4Navalny also pointed out that following the Yandex strikes, a group of known Russian disinformation accounts, a network known as Matroyshka, were still active on Tuesday. (Matroyshka is also the network that recently posted fake videos bearing Mother Jones’ name and featuring fake quotes from our DC editor David Corn.) 

It’s not entirely surprising that the bot takedown story spread so far and fast, even in the absence of evidence. “We live in an information environment in which people are rightly increasingly suspicious of artificial participation, even as we are increasingly unable to measure it,” DiResta told me. 

There’s also clear evidence that people have a difficult time separating authentic from inauthentic activity online, even as they overestimate their own ability to do so. A 2023 study from media scholars at Indiana University found that participants tended to “to overestimate the prevalence of bots and see others as more vulnerable to bot influence than themselves.”

While public perceptions on bots are understudied, the researchers added, “A consensus on the definition of social bots is lacking; people appear to use the term to refer to a variety of entities, from fake profiles and spammers to fully-automated accounts. This ambiguity provides social media users with a scapegoat for their unpleasant online experiences. For example, one may reject accounts with opposing political views by labeling them as bot.”

All of this, then, led to an environment where a claim was as viral as it was unproven. 

“The ‘Russian bot takedown’ story is satisfying because people are tired of suspected manipulation,” DiResta added. “And the idea that a Ukrainian strike silenced fake accounts is a compelling story of victory against a common enemy. It also supports the comforting belief that your political enemies aren’t even real.”

But however tempting it is, now isn’t the time to believe in fiction, even if it feels like fact.