Forged TV News Clips Spread False Senate-Race Claims Before November Vote

A real Cleveland television broadcast was recut into a fabricated report about Ohio Senate candidate Sherrod Brown, while other manipulated political videos targeted Senate races in Texas and Maine. With the November 3 congressional election approaching, the cases expose a way to put false claims into familiar news formats without establishing who made every clip.

The Ohio recording was not an ordinary political attack ad. It borrowed footage of a recognizable television anchor, kept elements of the original broadcast and replaced what viewers would hear. The source footage was authentic; the purported report was not. The distinction matters because a viewer encountering only the altered version could mistake fabricated allegations for reporting by an established broadcaster.

The Newscast That Was Rewritten

In an October 28 examination of the Cleveland 19 video, Lead Stories compared the circulating clip with a genuine October 21 segment. The original discussed an Ohio gubernatorial campaign advertisement. The fabricated version diverted the same opening footage into a false report involving Brown. The television station was being impersonated, not identified as the producer of the fake.

Lead Stories also tested nine seconds of the altered audio with a commercial detection tool, which estimated an 88.4% likelihood of AI-generated speech in that sample. That estimate supports the finding of manipulation; it does not identify the people who assembled or posted the clip. The direct comparison with the original broadcast is a separate and more concrete basis for identifying the substitution.

The video also borrowed the names and images of real people. In Associated Press interviews with an affected former Brown staff member, Haya Akbik rejected claims attributed to her as “completely made up.” PolitiFact separately reported another former staff member's denial. People used to lend credibility to the story disputed its contents. Their denials are central to assessing the material, not secondary reactions to an established allegation.

The pattern crossed state lines but did not take the same form everywhere. The AP documented an apparently manipulated Texas video that presented a purported account from a former student of Senate candidate James Talarico. In Maine, PolitiFact examined an altered WABI-TV report involving Senate candidate Troy Jackson. Two of the cases reused recognizable local-news coverage, while the Texas material impersonated an individual.

Researchers Link a Network, Not Every Operator

Researchers cited in PolitiFact's cross-state investigation link the material to Storm-1516, a Russia-associated influence network. Clemson University researcher Darren Linvill and other specialists pointed to similarities in format and amplification patterns. The connection is an attributed research assessment, not a disclosed forensic chain establishing Russian government direction of each video or the identities of the people who created them.

That uncertainty does not erase the verified manipulations. The Cleveland footage can be compared with the authentic newscast, and people represented in the videos have rejected statements assigned to them. But the evidence available here does not establish a complete distribution network, precisely how its accounts operated together or how many copies remain online. The production and amplification chain is only partly visible.

The apparent mechanism is straightforward: reuse familiar footage or a real person's identity, substitute invented content, then rely on social platforms to distribute it. In these cases, the danger is confusion about the source of a claim, not evidence that voting machines were compromised or ballots were changed. It remains a potential information-system effect rather than a measured electoral result.

Removal Has Not Established the Reach

The AP reported that X removed initial posts and suspended accounts associated with two videos, while copies were still circulating when it reported on October 27. Removing an original post does not establish that every copy has vanished. At the same time, continued circulation does not establish how many distinct U.S. voters watched or believed the material.

There is also countervailing evidence about scale. Linvill told PolitiFact that engagement was lower than in the 2024 cycle. That is a researcher's comparison, not a complete audience audit. Online views and reposts cannot be treated as verified voter persuasion, and the source record offers no demonstration that the clips shifted turnout or candidate preferences.

For broader context, NewsGuard's October 29 midterm misinformation tracker counted 37 debunked claims and 465 documented appearances across monitored online channels. Those figures describe NewsGuard's wider sample of election misinformation, not the reach of Storm-1516 or of these particular videos.

A separate October 16 FBI-CISA advisory warned election and campaign officials about impersonation, deceptive digital communications and synthetic audio or video. It did not assign responsibility for the Brown, Talarico or Jackson clips. The federal warning supplies context for the method, not official confirmation of the researchers' attribution.

The November 3 election date supplies the next hard deadline. What remains unknown is how many unique voters encountered the fabrications, who produced each one and whether exposure changed any vote. There is no demonstrated effect on election results. The verified record is narrower: forged content circulated, initial posts were removed, and the remaining human reach is unmeasured.