Russian propaganda is using deepfakes, sham websites and social media swarms

A deepfake video of a State Department official falsely claiming a Russian city is a legitimate target for Ukrainian strikes using U.S. weapons.

Pro-Russia social media accounts amplifying stories about divisive political topics such as immigration and campus protests over the war in Gaza.

Sham news sites spoofing real publications or posing as legitimate-sounding outlets with names like D.C. Weekly, the Boston Times and Election Watch.

Russian propaganda is ramping up in a busy global election year, targeting American voters as well as elections in Europe and the Paris Olympics, according to intelligence officials, internet researchers and tech companies.

“Russia remains the most active foreign threat to our elections,” Director of National Intelligence Avril Haines told senators last month at a briefing about election risks.

Influence operations linked to Russia take aim at a disparate range of targets and subjects around the world. But their hallmarks are consistent: attempting to erode support for Ukraine, discrediting democratic institutions and officials, seizing on existing political divides and harnessing new artificial intelligence tools.

"They're often producing narratives that feel like they're throwing spaghetti at a wall," said Andy Carvin, managing editor at the Atlantic Council's Digital Forensic Research Lab, which tracks online information operations. "If they can get more people on the internet arguing with each other or trusting each other less, then in some ways their job is done."

Some efforts have been linked directly to the Kremlin, including a network of fake accounts and phony news websites given the name Doppelganger, whose operators have been sanctioned by both the U.S. and the European Union.

The origins of others are still unknown, such as the fabricated video of State Department spokesman Matthew Miller, in which reporters' questions and Miller's response about U.S. policy in the Ukraine war were faked, likely with the help of artificial intelligence. The video circulated on Russian Telegram channels and was picked up by Russian state media and government officials, according to The New York Times.

Russia employed tactics such as exacerbating existing divisive issues and creating fake accounts posing as Americans in its 2016 and 2020 election-meddling efforts, researchers say. Since Russia invaded Ukraine in early 2022, discrediting Ukraine and amplifying voices in the U.S. and other countries that oppose aid to Ukraine and support for NATO has become a dominant theme of the Kremlin's efforts.

"What you can see is they are referencing politics in a certain country, and they are generally tying that to what is going on in Ukraine. The underlying message is, 'Here's why people should not support Ukraine,'" said Ben Nimmo, principal investigator on OpenAI’s intelligence and investigations team, who previously led global threat intelligence at Facebook's owner, Meta.

Fake accounts, phony websites

The Kremlin relies on what Haines called "a vast multimedia influence apparatus, which consists of its intelligence services, cyber-actors, state media proxies and social media trolls" to pump out propaganda, launder fake and misleading news articles and circulate conspiracy theories.

Since the invasion of Ukraine, the European Union has banned Russian media outlets including RT, Sputnik, Voice of Europe and RIA Novosti from publishing or broadcasting within the bloc. That hasn’t stopped RT articles from proliferating across hundreds of other websites widely available in Europe, according to a recent report from the German Marshall Fund of the United States, the University of Amsterdam and the Institute for Strategic Dialogue.

"We discovered RT articles reposted to third-party websites targeting audiences from Iraq to Ethiopia to New Zealand, often without any indication that the content was sourced from a Russian propaganda outlet," the researchers wrote.

Perhaps the most persistent and prevalent Russian online operation is Doppelganger. First identified by researchers at the EU DisinfoLab in 2022, the campaign has impersonated news outlets including the U.K.'s The Guardian, Germany's Der Spiegel, The Washington Post and Fox News, and it has posed as NATO, the Polish and Ukrainian governments, the German police and the French Foreign Ministry.

In addition to operating fake accounts and phony websites, the operation purchased Facebook ads targeting French and German audiences with messages about aid to Ukraine, farmers' protests and the war in Gaza, according to the European nonprofit AI Forensics.

Doppelganger has also set its sights on the Paris Olympics, Microsoft said in a report this week. It used fake French-language news sites to push claims of corruption in the Games' organizing body and to warn of potential violence.

In March, the U.S. Treasury sanctioned two Russian companies identified as being behind Doppelganger — Social Design Agency and Structura — as well as their founders, saying they carried out the campaign "at the direction of the Russian Presidential Administration."

The misinformation-tracking company NewsGuard has connected a separate network of 167 websites "masquerading as independent local news publishers in the U.S." to a former deputy sheriff from Florida who now lives in Moscow.

Using AI tools to create propaganda

The volume of posts, articles and websites that Russian-linked operations produce is being boosted by artificial intelligence — another new factor that sets 2024 apart from previous election cycles.

Covert influence campaigns based in Russia, as well as in China, Iran and Israel, have begun using AI in their attempts to manipulate public opinion and shape politics, according to recent reports from OpenAI, Meta and Microsoft.

A Russian operation that Microsoft calls Storm-1679 used AI to fake actor Tom Cruise's voice narrating a phony Netflix documentary disparaging the International Olympic Committee.

According to OpenAI, Doppelganger has used its AI tools, which include ChatGPT, to translate articles into other languages and generate social media posts and comments. Another Russian operation, dubbed Bad Grammar, used AI to debug code for a program that automatically posted on Telegram.

The question remains: How effective are Russia's attempts to influence public opinion and democratic elections?

Many online operations that have been publicly identified haven't reached large audiences of real people, researchers say, and AI hasn't made them any more convincing — at least not yet.

"It's absolutely true that when you look at an individual campaign, it's just as likely as not that it hasn't had a huge amount of influence, which is why Russia just does it again and again, or in a different form, or targeting a different group," the Digital Forensic Research Lab's Carvin said. "It's almost like producing cheaply manufactured goods and just getting it out there in the world, hoping that maybe one particular gadget ends up becoming the popular toy of the season, even if the others completely fail."

Many researchers who study disinformation warn against seeing the hand of Russia as an all-powerful puppeteer, especially since so much of what its mouthpieces amplify is homegrown.

"Any potential narrative that's being argued in a given political environment is fodder for Russian operations — which in itself can sound a little crazy and conspiracy-ish," Carvin said. "And in some ways you risk creating a … situation where absolutely everything that's happening online is all Russia's fault."

But, he added, "at the same time, Russia has a lot of resources at its disposal and it's willing to experiment in different ways to see which things stick. … Why not try all of the above and see where it takes you?"