Election Disinformation: Russian-Linked AI Videos Target Two Senate Races
AI-generated clips misuse real names and images to attack Democratic candidates in Ohio and Texas days before midterm voting concludes.
By Marina Weber ·

Russian-linked social media accounts have deployed artificial intelligence videos to attack two Democratic Senate candidates in the final days of the U.S. midterm elections. The clips target Sherrod Brown in Ohio and James Talarico in Texas, using the names and photographs of real individuals to fabricate negative narratives. Experts identify the material as consistent with the tactics of a specific Russian disinformation network.
The video targeting Ohio candidate Sherrod Brown mimics a television newscast. It features images of Haya Akbik, who interned for Brown, alongside other former staff members. Akbik stated to The Associated Press that the claims attributed to her in the clip are completely made up. She noted that the fabrication harms both her reputation and the senator. The clip appears to contain elements of AI manipulation.
In Texas, a similar video targets candidate James Talarico. The footage is framed as a direct-to-camera confession from a former student of Talarico. JT Ennis, a spokesperson for Talarico, said in a statement that powerful figures are slandering the candidate with AI deepfakes and foreign bots. The goal, according to Ennis, is to protect a broken and corrupt political system.
Darren Linvill, a professor of media forensics at Clemson University, identified the likely source of these videos. He found that the accounts sharing the new material had previously distributed content from Storm-1516. This Russian network targeted former Vice President Kamala Harris and Minnesota Governor Tim Walz in the 2024 election cycle. Linvill concluded that the new clips are likely products of the same network.
Hany Farid, a professor at Dartmouth College who studies artificial intelligence deepfakes, analyzed the technical aspects of the videos. He determined that the clip targeting Talarico is likely generated by AI and fits the pattern of disinformation released close to Election Day. The video targeting Brown, Farid noted, is a fairly low-tech fake but still relies on AI manipulation. His analysis suggests both videos were either generated or altered using the technology.
The social media platform X removed the initial posts containing the videos and suspended the user accounts for rule violations. Copies of the content remained accessible online as of Tuesday, with some versions accumulating hundreds of shares. This persistence illustrates the challenge of containing disinformation in the final days of an election cycle.
Haya Akbik said the incident underscores the need for Congress to regulate what she calls AI slop on the internet. She described the impact on her personal brand as damaging. The misuse of her name and image serves as a concrete example of how foreign adversaries exploit real people to influence American voters.
The U.S. Cybersecurity and Infrastructure Security Agency has identified Russian disinformation videos in past election cycles. The agency did not respond to a request for comment regarding these new clips. The Russian Embassy in Washington also did not immediately respond to requests for comment. The embassy has historically denied interfering in U.S. elections.
The use of AI in these campaigns demonstrates a familiar playbook by foreign adversaries. The technology allows for the rapid creation of damaging information that can appear highly realistic. As voting concludes in the final stretch, the spread of these fabricated narratives aims to influence which party controls Congress next year. The specific targeting of individual staff members and former students adds a layer of personal credibility to the false claims.
The videos exploit the final window of voter attention before ballots are cast. By using recognizable faces and names, the clips aim to bypass the skepticism voters might apply to anonymous accounts. The technical analysis by Farid and Linvill confirms the presence of AI tools in the production process. The suspension of accounts on X does not remove the content from the wider internet, allowing the disinformation to continue circulating in the critical period before election day.
Source: KPRC-TV





