Texas voters heading into the March 3 primary are getting a crash course in artificial intelligence, whether they asked for it or not. Campaigns across the state are cranking out AI-built attack ads that mock rivals, sharpen partisan hits, and lean into parody. The output ranges from goofy cartoon spoofs to slick deepfakes, including an AI spot that shows Sen. John Cornyn two-stepping with Rep. Jasmine Crockett and a separate fake video targeting Rep. Wesley Hunt. Voters, campaigns, and watchdogs say the surge is testing the limits of current rules and raising a basic question: can people still tell what is real in their feeds.
Paxton's Two Step At Cornyn
In mid January, Attorney General Ken Paxton’s campaign dropped a short AI-made ad that pasted synthetic clips of Cornyn dancing with Crockett onto real interview footage and news segments, then pushed it across social media. As reported by the Houston Chronicle, the message was not subtle. Cornyn’s record of working with Democrats was framed as a political liability for conservative primary voters.
The ad does include an on-screen note that labels some imagery as satire or AI-generated, but viewers and reporters quickly pointed to visual tells that gave the trick away anyway, including strange body movement and blurry, unstable backgrounds.
Democrats And Down Ballot Parodies
The Paxton spot is not an outlier. A review of recent Facebook and YouTube ads finds AI tools popping up all over the ballot, from graphics that caricature state Rep. James Talarico to parody videos that turn public figures into babbling babies.
According to The Texas Tribune, some candidates and parody channels have been upfront about using synthetic media, while others have left viewers guessing about what is authentic and what was cooked up by a model.
How Researchers And Tools Spot The Fakes
Specialized detection systems are already tagging some of the biggest clips as almost certainly fake. The Tribune reported that Hive Moderation and Google’s Gemini both put the odds at 99 percent that a Cornyn-posted video mocking Rep. Wesley Hunt was AI-generated.
“AI-generated videos and images will, in the long term, desensitize voters to the importance of factually accurate information,” Ernesto Calvo, a computational social science researcher, told The Texas Tribune. Angie Holan of the International Fact Checking Network warned that audiences often confuse parody with reality, so even content that is labeled or obviously satirical can mislead once it starts to spread at speed.
Where Lawmakers Hit Pause
State lawmakers tried to put some guardrails in place before the primary. A House bill that would have required political ads using AI or substantially altered images to carry clear labels passed the Texas House in 2025 but stalled out in the Senate.
The fight over that proposal, detailed by the Dallas Morning News, pitted transparency advocates against critics who argued the bill could chill satire and ordinary online speech. With that effort blocked, digital platforms and individual users are largely on their own to sort synthetic content during the 2026 cycle.
What Texans Can Watch For
Local coverage and digital forensics guides are offering a basic survival kit. Voters are urged to check for jerky or unnatural motion, lighting that does not quite match, background text that seems off, and crowds that dissolve into smudges when you look closely.
Hoodline’s local reporting flagged early spoofed mailers and images in Austin as a warning about how quickly fabricated material can race through communities, and the Houston Chronicle noted that some campaigns are pairing AI clips with disclaimers or production notes, a small hint that viewers should pause and hunt for original sources before sharing.
Experts say the most reliable move is old-fashioned verification. Look to see whether a piece of footage appears anywhere else in a credible, time-stamped context before taking it at face value. With the March 3 primary closing in and AI tools now a routine part of digital campaigning, the outcome may depend as much on voters’ media savvy as on the messages themselves. Texans who slow down, double-check, and watch for obvious synthetic tells are the ones most likely to keep the fakes from shaping the real results.
