Russian disinformation operators are running an AI-powered influence campaign against Democrats ahead of the November 3 midterm elections, flooding X with fabricated videos that put fake statements in the mouths of celebrities, according to a Bloomberg investigation published October 16 — and researchers who track the operation say it is only escalating as election day nears.
 |
| The Kremlin and Red Square in Moscow, Russia. (File photo) |
The campaign relies on two tools: artificial intelligence to manufacture convincing fake videos, and armies of automated bot accounts to inflate their reach. According to Bloomberg, the AI-generated clips — featuring the likenesses of well-known actors and musicians including Robert De Niro, Julia Roberts and Cher, appearing to attack Democrats — began appearing on X in early September. Some urge viewers to join a hashtag campaign called #AllDemocratsAreCriminals. Scores of other clips make false allegations of corruption and impropriety aimed at Democratic candidates.
How the Bot Army Works
The mechanics of the operation are as important as the fakes themselves. Bloomberg reports that the videos are posted by accounts with almost no real followers — and then swarmed by bots that inflate view counts and engagement, gaming recommendation algorithms so the material gets pushed in front of real users.
Social media researchers have attributed the posts to a Russian state-backed operation known as Matryoshka, or Operation Overload — a network that has been running influence operations for years but has adapted its tactics for the AI era. The name captures its method: nested layers of fake accounts, fake content and fake engagement, each amplifying the others.
The content is not limited to one race or one office. According to the reporting, the fabricated clips are aimed at Democratic candidates across the country, mixing fake celebrity endorsements of anti-Democratic narratives with bogus news branding — including one clip carrying an ABC logo that made a false claim about a Texas Democrat, according to reports.
The New York Times and Intelligence View
The campaign has also drawn coverage from The New York Times, which, alongside Bloomberg, reported on the operation on October 16. According to Newsmax's account of that reporting, U.S. intelligence officials believe Russia's broader objective is to exploit divisions among Americans and create distrust around the election itself, rather than specifically helping one political party win.
Still, one prominent theme in the documented campaign is the targeting of Democrats, with fabricated videos attacking Democratic candidates and amplifying false allegations involving corruption and other politically divisive issues, Bloomberg reported. The apparent tension between the two goals — sowing chaos generally and targeting Democrats specifically — is itself a known feature of Russian influence operations, which researchers say often play both sides to deepen polarization.
The stakes are unusually high this cycle. With control of the House and Senate on the line on November 3 — all 435 House seats and 35 Senate seats are up for election — even a small shift in voter sentiment or confidence could matter.
 |
| The U.S. Capitol in Washington, D.C. (File photo) |
What This Means
The operation marks a new chapter in election interference: the industrialization of fakes. In 2016, Russian influence operations largely relied on stolen documents, fake grassroots pages and human-run troll farms. Now, according to the Bloomberg reporting, AI allows the same kinds of operations to manufacture convincing political content faster and more cheaply than traditional propaganda — and bot networks can distribute it internationally within hours.
Researchers cited by AFP warned that the difficulty is no longer creating the fakes — it is detecting them at scale. A clip that uses a celebrity's cloned voice, authentic footage taken out of context and the logo of a trusted news outlet can fool even careful viewers scrolling quickly through a feed.
The campaign also exposes a persistent vulnerability in how social platforms measure popularity. By manufacturing views and engagement, bot armies can push disinformation into recommendation systems designed to reward whatever is trending — effectively laundering propaganda through the platform's own algorithms.
 |
| A voting booth in the United States. (File photo) |
What Happens Next
Researchers say additional material is likely to appear as the November 3 elections approach, with the volume of fakes expected to rise in the final stretch of the campaign. That pattern matches past cycles, where influence operations intensified in the weeks before voting.
Whether American institutions respond more aggressively remains an open question. The Bloomberg investigation landed against a backdrop of renewed debate over platform accountability and election security funding, and Democratic lawmakers have already called for stronger measures against foreign interference this cycle.
For voters, the immediate defense is skepticism: extraordinary claims from celebrity accounts or unfamiliar news brands deserve a second look before a share. But as AI tools grow more capable and bot networks more sophisticated, that advice is doing more work than ever before.
Post a Comment