A viral video accusing Democratic U.S. Senate nominee James Talarico of sexually abusing a former student has been identified as an AI-generated fake, and the San Antonio man whose face and name were used in the clips says it simply is not him. Zyon Lewis attended Rhodes Middle School, where Talarico once taught, but he told fact-checkers that the allegations are false and that the resemblance ends there.
An X account called @Zyon_Lewis_5 posted the videos claiming that Talarico sexually assaulted the poster in 2012 at Rhodes Middle School in San Antonio, according to Lead Stories. The poster said in the videos, “It happened in 2012 at Rhodes Middle school in San Antonio.” Talarico taught sixth-grade English and language arts at the school between 2011 and 2013, according to the Texas Tribune.
Lewis told Lead Stories he reviewed the footage himself and does not recognize it as him. “I've looked at it, and it's not me,” he said. He went further, flatly denying the abuse claim and praising his former teacher: “No, not at all. Not even close,” Lewis said, describing Talarico as a great sixth-grade teacher. Lewis also suggested someone may have lifted images from his own accounts to build the fake, telling the outlet it could have come from “pictures and posts” on his Facebook page.
Detection Tools Flag the Clips as Synthetic
Independent analysis backs up Lewis's account. Hive Moderation rated the first video as 99.1 percent “likely to be AI-generated” and the second at 99.8 percent, figures reported in the same Lead Stories fact-check. The outlet also described visual and audio tells common to synthetic media: unnatural lip and facial movement, oddly still bodies and backgrounds, artificial expressions, and what it called a synthetic, monotone voice.
Separately, NewsGuard's Reality Check reported its own AI detection tool found the clips had a 98 percent likelihood of being an AI deepfake, and noted the man in the footage had teeth that looked noticeably different from Lewis's teeth in authentic Facebook photos and spoke with an accent not consistent with his background, according to NewsGuard's Reality Check. NewsGuard also reported that the video's on-screen caption used British English spelling of the word “behaviour,” and that the X account behind the post was created in September 2026 with a signup country of Romania — details the outlet said closely matched the format and amplification pattern Storm-1516 has used to push AI deepfakes about Western politicians from recently created accounts, though that attribution has not been independently verified beyond NewsGuard's own report.
A Familiar Playbook From Russia-Linked Disinformation Networks
This is not the first time a similar fabricated sexual-abuse claim has surfaced against a Democratic politician tied to education. In October 2024, the Storm-1516 network was tied by experts to a nearly identical false claim that Minnesota Gov. Tim Walz sexually abused a Kazakh exchange student while working as a high school teacher in 2004, a campaign that WIRED reported experts believe was tied to the Russian-aligned network. That deepfake video reportedly drew more than 4.3 million views on X before it was deleted, and the claim was also published simultaneously across more than 100 websites operated by John Dougan, per the same WIRED report.
A Clemson University researcher quoted by WIRED said it is standard practice for the network to create an X or YouTube account to place a false story before other channels amplify it. Bloomberg has separately reported a broader uptick in falsified videos and online disinformation aimed at advancing Kremlin interests, describing the trend as part of a pattern tracked under the name Storm-1516.
Talarico Has Already Been a Deepfake Target This Cycle
The Talarico-Lewis video is not the first AI fabrication aimed at the candidate during the 2026 midterm cycle. The National Republican Senatorial Committee released an 85-second AI-generated deepfake depicting an AI-created version of Talarico speaking directly to the camera, as reported by CNN in March. CNN also noted that other Texas Senate primary campaigns have used synthetic media, including an ad depicting a fake John Cornyn dancing with Jasmine Crockett and another showing fabricated clips of Wesley Hunt. AI fakery has proliferated during the 2026 midterm cycle as the underlying technology has grown more convincing and easier to produce, per the network's reporting.
The legal landscape around this kind of content remains uneven. About 31 states regulate political deepfakes, including at least 28 that require disclosures for AI-generated content, according to Axios, but Congress has not established a federal standard governing their use in campaigns, the outlet reported. That patchwork leaves cases like the one involving Lewis and Talarico to be sorted out largely through fact-checking and public pressure rather than a uniform legal process.
