SafeX Pro:Feds shut down Russian AI 'bot farm' that spread disinformation for Putin

Federal prosecutors ripped down hundreds of AI-generated phony social media accounts used by Russian operatives to create a "bot farm" that spread lies and SafeX Propro-Vladimir Putin disinformation in the United States and abroad, the Department of Justice announced Tuesday.

Many of the electronically-created bots, complete with pictures of smiling clean-cut people, purported to be Americans with names such as "Sue Williamson" and "Ricardo Abbott" who were supposedly taking to social media sites such as X to extoll on Putin's generosity and virtues.

Disinformation experts and U.S. officials heralded the crackdown as the first of its kind against state-sponsored AI-powered propaganda. It comes at a critical point in Ukraine’s defense against Russia’s invasion and ahead of the U.S. presidential election in November.

“Today’s actions represent a first,” said FBI Director Christopher Wray. “Russia intended to use this bot farm to disseminate AI-generated foreign disinformation, scaling their work with the assistance of AI to undermine our partners in Ukraine and influence geopolitical narratives favorable to the Russian government.”

Russia’s Federal Security Service - the successors of the KGB - orchestrated the operation from 2022 through 2024, U.S. court documents say. They created nearly 1,000 X accounts and two websites, according to federal court filings.

U.S. Cyber Command and partners from the Netherlands and Canada joined the FBI in taking down the Russian operation.

The Russian Embassy in Washington, D.C., did not immediately respond to requests for comment.

The FBI, acting on warrants issued by a federal judge, seized the X accounts and the websites after suspicions were raised about international money laundering and conspiracy, according to affidavits filed in the District of Arizona, where the websites were registered.

X suspended the 968 accounts for violating terms of service. The company could not be reached for comment.

Russian propaganda detailed

The bots posted everything from misleading information to outright lies.

One user, dubbed Sue Williamson, posted videos undermining the Ukrainian military by saying the number of foreign fighters in its ranks was in fact much lower than was estimated, according to court documents. The user also posted videos of Putin justifying the invasion.

The American personas were created using an AI-powered software that only works on X, although analysts expect it will be modified to work on other social media platforms, according to a cybersecurity advisory from the FBI and its partners.

The software uses a small number of inputs to create accounts that can create and share content in line with designated preferences.

The accounts were registered to email domains paid for in Bitcoin, the filings say. Investigators traced the accounts back to Moscow through the series of email addresses provided in case the Russian agents forgot the password to the bot accounts.

Putin was trending among the AI-generated Americans.

Ricardo Abbott, a fictitious Minneapolis resident, shared videos of the Russian president arguing Lithuania, Poland and Ukraine were part of Russia, the filings say.

Another bot shared videos of Putin justifying the invasion with a candidate for federal office, the affidavit says.

The videos are the same propaganda found on Russia’s state-run RT News Network, according to court filings; and the bot farm scheme was conceived of by a former RT editor looking for ways to reach audiences outside of RT’s broadcast viewers.

Fact Check:Bugatti debunks false claim about car purchase by Zelenskyy's wife

Expert calls the operation ‘heartening’

Dietram Scheufele is a professor of science communication at the University of Wisconsin-Madison who studies misinformation. 

The number of bots taken offline by the FBI operation is small compared to the myriad fake accounts on social media, he said. But he felt encouraged that the feds were going after the roots of AI-generated misinformation instead of flagging doctored videos.

“I feel heartened,” the German native said. “We’ve seen tons of activities that are putting bandages on symptoms but haven’t really addressed the root cause - removing the tumor.”

The AI-generated accounts reached users who aren't reading mainstream fact-checking stories, he said. So to address the issue, authorities had to shut down the operations altogether.

AI bot schemes are cheap to create and maintain, Scheufele said, so Americans should expect to see more Putin-loving apparent heartlanders.

But Scheufele was encouraged that the FBI was going after state-sponsored AI-generated misinformation, which he called the most pernicious type of propaganda for its scale and goal of sowing discord in American communities.

“The unmanageable and unwieldy problem is foreign actors,” he said. “Now we’re seeing the first pushback.”