The Fort Bend County judge who is also running to keep his seat is speaking out after a video on his campaign’s social media pages was identified as using artificial intelligence-generated images of Black voters.
The video showed four people praising Judge Daniel Wong. After it was identified as using AI-generated imagery, the campaign removed it from its platforms. Wong claims his team had nothing to do with the video, but he also called out his opponent, Democrat Dexter McCoy, for criticizing the support shown in it.
The incident is raising broader questions about AI-generated content in political campaigns and how voters can protect themselves from being misled.
Chris Bronk, a professor of public affairs at the University of Houston, says this kind of content is everywhere on social media and spotting it is getting harder by the day.
“There are a number of cues,” Bronk said. “But even the best among us at looking at this have largely given up in knowing for sure and certain that things are made by AI, to the degree now that governments are talking about putting watermarks in any video made by AI that can’t be removed.”
When asked whether someone could grab a photo from Instagram, Facebook or TikTok and use it to create an AI-generated video without permission, Bronk said the technology makes it very possible.
“People can make the videos, that’s definitely doable,” Bronk said. “The question is what will be depicted in the video and whether that is in some way a violation of law. That’s the question.”
So what should viewers look for when trying to spot a fake? Some of the telltale signs include visual inconsistencies, objects or people appearing and disappearing, text or detail errors, unrealistic textures, unnatural or stiff motion, people who don’t blink and flat, robotic tones.
But Bronk says even those red flags are becoming less reliable as AI continues to improve.
“You’re covering all the kind of glitchy little odd behaviors that we see,” Bronk said. “But AI is getting smarter every day, so spotting the fake won’t always be simple.”
As for what to do if you find yourself depicted in an AI-generated video without your consent, Bronk says it depends on the content.
“If a viewer has seen themselves depicted in a deepfake video that is compromising in some way, that is a ‘call the local police, do not stop, go’ kind of matter, that’s real,” Bronk said. “Anything in the not inflammatory space, if someone just depicted a video of me walking down the sidewalk and I said, ‘I don’t think that’s me.’ calling the police on that is not going to be easy.”
