# Why Did Russian Deepfake Videos Get 200,000 Views but Almost No Likes? > Bloomberg found Russia-linked Matryoshka clips hit about 200,000 views in an hour with almost no likes. - Canonical URL: https://www.socialsignalreview.info/blog/russian-deepfake-videos-fake-view-counts-engagement-signals/ - Publisher: Social Signal Review (https://www.socialsignalreview.info/) - Author: Luca Bianchi - Published: 2026-10-27 - Updated: 2026-10-27 - Topics: Social and Community Signals, Reviews and Reputation, AI Brand Monitoring ## Key takeaways - Bloomberg reported some Matryoshka clips reached about 200,000 views in the first hour, then plateaued, with almost no likes or reposts. - Amplifying accounts were often dormant for years with few followers, per Antibot4Navalny data cited by Kyiv Post. - Deepfakes of De Niro, Roberts and Cher urging people not to vote for Democrats were posted on X under real news outlets' logos. - Researchers at Graphika, Alethea and Clemson linked videos baselessly accusing James Talarico to Storm-1516, per NBC News. - Warner and Himes cited a classified US intelligence report; no proven attempts to hack election systems. ## Why Did Russian Deepfake Videos Get 200,000 Views but Almost No Likes? Because the views were not real. Bloomberg reported that bots linked to Russia's Matryoshka operation inflated view counts on deepfake celebrity clips and fake "news" videos aimed at the 2026 US midterms. Some reached about 200,000 views in the first hour, then stalled, with almost no likes or reposts. That gap between reach and response is a signal anyone can learn to read. ## What happened in this campaign? Researchers and US lawmakers describe a Russia-linked effort to discredit congressional candidates with AI-generated video before the November 3 vote. The main threads, as reported: - **The intelligence claim:** Sen. Mark Warner and Rep. Jim Himes, senior Democrats on the intelligence committees, said a classified US report shows Russia using AI tools to undermine candidates and trust in the election, [Politico reported on October 21](https://www.politico.com/news/2026/10/21/himes-warner-russia-election-interference-report-01130967). They believe Democrats and pro-Ukraine Republicans could be targeted. There are no proven attempts to hack election systems. - **The celebrity deepfakes:** AI clips made Robert De Niro, Julia Roberts and Cher appear to urge people not to vote for Democrats. Researchers attributed them to the Matryoshka network, according to [Meduza's account](https://meduza.io/en/feature/2026/10/23/for-weeks-deepfake-celebrities-have-been-telling-americans-not-to-vote-for-democrats-in-the-midterm-elections-u-s-intelligence-says-russia-is-behind-it). - **The borrowed credibility:** The clips were posted on X under the logos of real news outlets. - **A second network:** [NBC News reported](https://www.nbcnews.com/tech/misinformation/russian-operatives-likely-spread-fake-videos-targeting-texas-senate-ra-rcna604923) that Graphika, Alethea and Clemson University researchers linked videos baselessly accusing Texas Democrat James Talarico to Storm-1516, a network researchers tie to Russian military intelligence. ## What does fake engagement actually look like? It looks like a big number with nothing behind it. [Bloomberg's investigation](https://www.bloomberg.com/news/features/2026-10-16/russian-bots-manipulate-view-counts-on-fake-videos-ahead-of-us-midterms) found a consistent pattern, and [Kyiv Post's summary](https://www.kyivpost.com/post/87423) adds forensic detail from the researcher group Antibot4Navalny: 1. **A vertical spike.** Views jumped to roughly 200,000 within the first hour. 2. **A flat line afterward.** The counts then stopped growing, which genuine attention rarely does. 3. **Missing reactions.** Likes and reposts were almost absent relative to views. 4. **Hollow amplifiers.** The accounts pushing the posts had often been dormant for years and had few followers. Each signal alone can have an innocent explanation. All four together describe a manufactured audience. ## How is real community response different? Real audiences leave traces beyond the view counter. When people actually care about a post, the numbers move together and keep moving: - **Growth is gradual.** Organic engagement usually builds over hours as people share to their own networks, rather than arriving in a single burst. - **Reactions scale with reach.** More views bring proportionally more likes, replies, quotes and reposts. - **Sharers have histories.** Genuine amplifiers have years of varied posting and real followers, not long silences broken by one political clip. The Talarico case shows how this mixes in practice. Clemson's Darren Linvill told Straight Arrow News that about 15 marketing bots shared the post immediately, and paid influencers then amplified it. Once real people start resharing, the line between fake and organic reach gets blurry. That is why early signals matter. ## Why does an inflated view count work at all? Because people treat popularity as proof. A clip showing hundreds of thousands of views looks widely watched and, by implication, widely believed. A false claim borrows credibility from the counter, just as the Matryoshka clips borrowed it from news logos. Bloomberg's findings suggest the view number itself was part of the deception, not just a measure of it. ## What should reputation teams watch for? Reputation teams should watch for any mismatch between apparent reach and real response, especially around content using their name, logo or people. Specific things to monitor: - **Your marks in someone else's video.** Fake clips here carried real outlets' branding. One case logged by the [AI Incident Database](https://incidentdatabase.ai/cite/1715/) involved a Matryoshka deepfake falsely presented as a Mother Jones production; X removed it after the outlet complained. - **Ratio alerts.** Flag posts about your brand where views climb fast but likes, replies and shares stay near zero. - **Amplifier quality.** Check whether accounts pushing a narrative are new, long dormant or low follower. - **Plateau shapes.** A sudden spike followed by a flat line is worth investigating before anyone responds publicly. - **Platform actions.** X said on October 15 it had suspended more than 13,000 accounts tied to suspected Iranian, Russian and Chinese influence operations, so takedowns can change the picture quickly. Responding loudly to a fake that almost no one actually saw risks giving it new reach. Reading the engagement signals first helps teams decide whether to report it quietly, correct it publicly, or both. ## How can teams build this into routine monitoring? Make engagement quality a standing metric rather than something checked only during a crisis. Teams that already assess partners can apply the same logic: a [creator audit](https://markgrid.ai/product/creator-audit) examines whether an influencer's audience and engagement look authentic. That matters when, as Linvill described, paid influencers can carry a bot-seeded post into real feeds. Reputation now also lives in AI answers. [Model Share](https://markgrid.ai/product/model-share) tracks how AI models represent a brand, and the [GEO guide](https://markgrid.ai/geo-guide) explains how that visibility works. For the full timeline, attributions and a checklist for spotting deepfakes, see MarkGrid's [full breakdown of the Russian deepfake midterm campaign](https://markgrid.ai/resources/russian-deepfake-celebrities-midterms-how-to-spot). ## Frequently Asked Questions ### What is the clearest sign that engagement is fake? A large view count with almost no likes or reposts, especially after a sudden first-hour spike that then flattens out. Bloomberg found this pattern on Matryoshka posts. ### Who is behind the deepfake celebrity videos? Researchers attribute them to Matryoshka, a Russia-linked network also called Operation Overload. Separately, Democratic lawmakers say a classified US intelligence report describes a Russian AI-enabled campaign against candidates. ### Did the fake videos hack any election systems? No. Lawmakers say there are no proven attempts by Russia to hack election systems or campaign networks. ### Should a brand respond publicly to a fake video using its logo? Check the real engagement first. If few genuine people saw it, report it to the platform, as Mother Jones did with X. A public correction makes more sense if it is spreading organically. ### Can high engagement still be fake? Yes. Bots can seed a post and paid or unwitting real accounts can then amplify it, as researchers described in the Talarico case. Check who is sharing, not just how many. ## Sources - [Politico: Himes, Warner on Russia election interference report](https://www.politico.com/news/2026/10/21/himes-warner-russia-election-interference-report-01130967) (2026-10-21) - [Bloomberg: Russian bots manipulate view counts on fake videos ahead of US midterms](https://www.bloomberg.com/news/features/2026-10-16/russian-bots-manipulate-view-counts-on-fake-videos-ahead-of-us-midterms) (2026-10-16) - [Kyiv Post: Summary of Bloomberg Matryoshka investigation](https://www.kyivpost.com/post/87423) (2026-10-18) - [Meduza: Deepfake celebrities telling Americans not to vote for Democrats](https://meduza.io/en/feature/2026/10/23/for-weeks-deepfake-celebrities-have-been-telling-americans-not-to-vote-for-democrats-in-the-midterm-elections-u-s-intelligence-says-russia-is-behind-it) (2026-10-23) - [NBC News: Russian operatives likely spread fake videos targeting Texas Senate race](https://www.nbcnews.com/tech/misinformation/russian-operatives-likely-spread-fake-videos-targeting-texas-senate-ra-rcna604923) (2026-10-22) - [AI Incident Database: Incident 1715, Matryoshka Mother Jones deepfake](https://incidentdatabase.ai/cite/1715/) (2026-09-22) --- Cite as: Luca Bianchi. "Why Did Russian Deepfake Videos Get 200,000 Views but Almost No Likes?." Social Signal Review, 2026-10-27. https://www.socialsignalreview.info/blog/russian-deepfake-videos-fake-view-counts-engagement-signals/ HTML version: https://www.socialsignalreview.info/blog/russian-deepfake-videos-fake-view-counts-engagement-signals/ Markdown version: https://www.socialsignalreview.info/blog/russian-deepfake-videos-fake-view-counts-engagement-signals.md