Algorithms of Manipulation

As in previous years, the foundation of these operations consisted of thousands of fake profiles posing as ordinary users or page administrators with shared interests. Meta had repeatedly exposed and blocked such networks in the past. For example, in 2022 it uncovered the largest Russia-linked network since the start of the war: more than 1,600 fake accounts and pages that coordinated the spread of Kremlin war propaganda. This campaign, launched in May 2022, included over 60 fake news websites imitating well-known outlets (Spiegel, The Guardian, Bild, and others) in multiple languages. After publishing pro-Kremlin content (smearing Ukraine, discrediting refugees, claiming that sanctions were ineffective and only harmed the countries imposing them), the organizers widely shared links to these sites through a network of fake accounts on Facebook, Instagram, Twitter, and Telegram, and also promoted them through paid advertising.

Meta described this operation as a combination of technically complex elements (imitating media websites in seven languages required significant resources) and brute force (numerous fake profiles and advertising). This “hybrid” approach allowed the perpetrators to reach a wide audience relatively quickly: instead of spending time on organic follower growth, they immediately flooded the information space with a massive wave of identical messages. Such direct “frontal attacks” on the information environment are aimed more at short-term impact (“hit-and-run”) than at building long-term trust, but they can inflict substantial damage before being neutralized.

1,600

fake accounts and pages coordinating Kremlin war propaganda

Facebook’s ad targeting algorithms became tools for propagandists. Investigations by AI Forensics revealed that pro-Russian groups actively purchased political ads while circumventing oversight. Thousands of such ads were found from anonymous pages pushing Kremlin-aligned messaging in EU countries. Over 65% of this political content lacked mandatory labeling — meaning Facebook failed to mark it as political or disclose the sponsor. This points to vulnerabilities in Facebook’s ad moderation systems. Actors created throwaway pages to publish a single paid post, quickly reaching hundreds of thousands of users before deleting the page. “The page exists solely to pay for the ad, gain immediate reach, and then disappear — this is more effective than trying to grow a page over time,” explains AI Forensics researcher Paul Bouchot.

In addition, targeted advertising was used not only to inject ideas but also as an analytics tool: Russian operators monitored which messages resonated most, tracking user reactions in real time. One of the documents created by pro-Kremlin political strategist Ilya Gambashidze as part of the disinformation campaign known as the Good Old USA Project explicitly states: “Facebook targeting allows you to track audience reactions to distributed materials and adjust subsequent launches based on which group was most strongly affected.”

Thus, Facebook’s advertising platform served as a kind of testing ground for propaganda, where adversaries used A/B testing to select the most effective narratives. In the second half of 2022, Meta’s specialists managed to manually identify and neutralize malicious accounts that had not yet been removed by automated systems before the start of a full-scale investigation.

>65%

of political ads ran on Facebook without required labeling

Facebook’s feed algorithm prioritizes content likely to generate strong emotional reactions — anger, outrage, fear — to increase user engagement. Propagandists exploited this by posting highly provocative messages and images designed to spark conflict. The more users commented or shared (even in outrage), the more the algorithm promoted the post. For pro-Russian disinformation, this meant that highly polarizing content gained disproportionate visibility. Studies have shown that platforms optimized for engagement unintentionally accelerate the viral spread of lies and conspiracy theories.

Moreover, actors used artificial engagement boosting tactics: bot groups or controlled accounts would immediately like and comment on a target post to “boost” it in algorithmic rankings. In the aforementioned “Good Old USA” plan, it was noted that a “bot network” would moderate top-level discussions in comment sections to steer the narrative. Dozens of coordinated comments would appear under each targeted post, creating the illusion of widespread support or attacking opposing views. Regular users, seeing such “resonance,” might mistakenly believe the propaganda reflects mainstream opinion.

Russian operators created and maintained Facebook communities in advance — sometimes for months — posing as local interest groups — for example, Trump supporter groups in specific states (“Alabama for America the Great,” and similar). These groups were filled with neutral or thematically adjacent content that raised no suspicion. The plan was that at a decisive moment, these “sleeper cells” on the social network would be activated: administrators would abruptly switch to distributing campaign materials, turning groups with already established audiences into tools for influencing public opinion. Court documents in the Gambashidze case, declassified by the U.S. Department of Justice, revealed intentions to create 18 such cells in key U.S. swing states, so that at the critical moment they could work in unison to promote targeted narratives and links.

This tactic relies on Facebook’s built-in recommendation mechanisms: people in such groups begin sharing materials that can then spread through their friends’ feeds, while the algorithms may recommend joining popular groups to friends with similar interests, thereby increasing reach.

18 per state

“sleeper cells” were planned in every key U.S. swing state

Although this report focuses on Facebook, many operations were in fact cross-platform. Fake content first appeared on other platforms — for instance, AI-generated videos on YouTube or alleged “leaks” on Telegram — before being shared on Facebook by ordinary users. This way, propagandists leveraged the entire social media ecosystem: even if one platform banned their accounts, they still achieved reach via another. In many cases, Facebook became a secondary amplifier of disinformation that originated elsewhere — especially when it was shared by real users who had been misled. For example, fake videos of “Ukrainian soldiers threatening Trump” were originally disseminated through pro-Russian channels and then picked up by U.S. conspiracy groups on Facebook — vastly increasing their reach.

In this way, the Russian propaganda network on Facebook constitutes a complex structure of hundreds of fake accounts and pages acting in coordination to spread Kremlin-aligned narratives. These accounts, many of which interact with one another artificially, create the illusion of broad support for Russia and its policies, particularly in the context of the ongoing war in Ukraine. They comment on and repost each other’s content to boost engagement scores and maximize reach.

Many of these pages have thousands of followers and regularly post images and videos showcasing Russian military equipment and symbols — including the “Z” symbol associated with Russian armed forces. Pages like Vladimir Poutine Z illustrate how visual materials and national symbols are used to intensify pro-Russian sentiment.

The lack of adequate moderation for such posts may be linked to challenges in handling international content and the ambiguity of the rules. Since pro-Russian pages target an international audience, moderation algorithms do not always flag them as violating standards, especially if they avoid using explicit markers of disinformation. This gap allows pro-Russian narratives to spread freely despite numerous complaints from researchers and activists.

Facebook’s inadequate response leaves the platform vulnerable to foreign information operations that undermine support for Ukraine and promote pro-Russian sentiment among Western audiences. Without robust oversight and moderation, such content will continue to shape public perception of global affairs.

Given its extensive reach, Russia’s disinformation campaign on social media has a significant impact on public opinion in the U.S. and other Western countries. Through visual cues, symbols, and militaristic narratives — such as the “Z” symbol and imagery of Russian military equipment — Russian propaganda fosters a favorable image of the Russian armed forces and reinforces Kremlin policies. This content spreads via hundreds of fake accounts and pages that exploit Facebook’s algorithm to reach broad audiences, including users who never followed such pages.

According to a report by the U.S. Senate Intelligence Committee, Russia actively uses social media to promote militaristic propaganda, project strength, and appeal to Western audiences. Military imagery, such as tanks and missiles, creates a compelling narrative for patriotic users and helps mobilize pro-Russian sentiment — especially in the context of the war in Ukraine.

Other reports, such as the analysis by the Center for Strategic and International Studies (CSIS), highlight how Russian operations use bots and fake accounts to simulate public support. CSIS described how anti-Ukrainian narratives were spread using AI and automated accounts, creating the illusion of mass approval among U.S. users.

An in-depth study by our project on the role of social media in political propaganda also shows that social networks have become a key channel for spreading political propaganda, largely pushing traditional media into the background. Around two-thirds of the world’s population use social media, and more than half of adults — for example, in the United States — regularly get their news from these platforms. Platform algorithms (Facebook, Twitter/X, TikTok, Telegram, and others) facilitate the viral spread of content — research shows that false news travels faster on social networks than truthful information. Mechanisms such as recommendation feeds and interest-based groups create “echo chambers,” where users are exposed predominantly to information that reinforces their existing views.

Under these conditions, social media have become an extremely powerful tool for shaping public opinion — a fact actively exploited, including for state propaganda. They have turned into the most potent means of influencing minds, facilitating the spread of both information and disinformation.

The spread of pro-Russian narratives also contributes to “war fatigue” among Western audiences. Continuous exposure to posts showcasing the Russian military or presenting a positive image of Russia can create a false sense of legitimacy and strength — weakening support for Ukraine. This may lead Western societies, especially in the U.S., to question the rationale for continued support — ultimately weakening the international position of democratic allies and reducing financial and military aid to Ukraine.

Russian disinformation campaigns also undermine trust in Western democratic institutions, portraying them as incapable of adequately responding to information threats and protecting their citizens from external influence. According to a study by the Georgetown Journal of International Affairs (GJIA), Putin’s propaganda often focuses on portraying Western governments as ineffective and support for Ukraine as unnecessary.

Russian propaganda fosters a negative perception of Western institutions, which may influence — and perhaps is already influencing — the political preferences of American voters and contribute to strengthening pro-Russian sentiment within the United States.

The question of how to regulate the spread of foreign propaganda on American platforms is complex, as it requires balancing the protection of free speech with national security. In the case of Russian propaganda disseminated through Facebook, there is a possibility that pages promoting pro-Russian narratives may violate both U.S. laws and certain regulatory requirements governing specific business practices. For example, the Federal Trade Commission (FTC) and the U.S. Securities and Exchange Commission (SEC) have rules aimed at protecting users, ensuring transparency in business practices, and preventing manipulation.

Sources and References: