A pro-Trump political action committee released an AI-generated advertisement depicting Democratic candidate for the Texas Senate James Talarico wearing a dress [1].
The ad represents a growing trend of using deepfake technology to target political opponents with deceptive imagery. This specific campaign focuses on transgender-rights issues to influence voters in the Texas Senate race [1, 2].
The 15-second video [4] shows Talarico singing a parody of the song “My Favorite Things,” with lyrics centered on transgender children [1, 2]. The content was produced and aired on television in Texas by Citizens for Sanity, a pro-Trump PAC [2, 3].
Financial reports regarding the cost of the campaign vary. One report said that Citizens for Sanity spent six figures on the AI-generated campaign [3]. Another report said that the MAGA PAC dropped millions on the anti-trans AI advertisement [3].
Critics of the ad said the tactic is designed to push an anti-trans narrative through fabricated visuals [1, 2]. The controversy has led to calls for lawmakers to implement bans on the use of AI-generated content in political advertising to prevent the spread of misinformation [1, 2].
Talarico has been targeted by these deceptive tactics as part of a broader effort to frame his positions on transgender rights [1, 2]. The use of AI to place a candidate in a fabricated scenario, such as wearing a dress while singing, marks a shift toward more aggressive visual manipulation in state-level elections [1, 4].
“The 15-second video shows Talarico singing a parody of the song “My Favorite Things.””
The deployment of high-budget AI deepfakes in state senate races signals a transition from text-based misinformation to sophisticated visual deception. By using AI to create compromising or ridiculing imagery, political actors can bypass traditional fact-checking speeds, potentially altering voter perception before the authenticity of the media can be verified.
