Donald Trump has expressed concerns about the use of AI in the upcoming election.
The former president called AI “very dangerous” and accused his rival, Kamala Harris, of using it to inflate the attendance figures at her rallies.
“There was no one on the plane and she manipulated it with an AI,” he wrote, adding: “Creating a fake image constitutes election interference and she should be disqualified. Anyone who does that is cheating on anything!”
Despite Trump's claims, the crowd in the photo was real: A Harris-Waltz campaign official told Business Insider that the photo was taken by campaign staff and “was not altered in any way by AI.”
A few weeks after posting his accusation that Harris was using AI, the former president posted several apparently AI-generated meme images on Truth Social and X, featuring real people including Kamala Harris and Elon Musk.
He also re-shared to Truth Social a set of questionable images of Taylor Swift and her fans with pro-Trump messages. While two of the images appeared to depict real Trump supporters, others, including several photos of women wearing “Swifties for Trump” T-shirts, bore clear signs of being AI-generated.
Trump later denied that he created the images and reflected on the use of AI in general.
“We see it all the time,” Trump said of AI-generated content. “It's a little bit of a dangerous situation.”
Swift's photo has already garnered criticism and discussion online, which may be exactly what Trump wanted, but the former president's decision to post what appears to be AI-generated content without clearly labeling it could confuse some voters.
Virus misinformation
Experts have long worried that widespread use of AI could influence elections around the world.
Deepfakes, AI-generated text, and hyper-targeted ads are blurring the line between fact and fiction, making it difficult for voters to discern trustworthy information, and bad actors could use AI tools to spread deliberately realistic misinformation on social media at an unprecedented scale.
Public literacy regarding AI-generated content also varies: some people can easily spot fake content, while others are fooled by fake images and text.
“This content is being spread around without anyone properly fact-checking it,” Steg.AI CEO Eric Wengrowski told Business Insider, “and it comes from a former president.”
Wengrowski said it's not unethical for politicians to share AI-generated content for the purposes of humor or satire, but public figures should ensure the content is AI-generated.
“I find it unethical for politicians to share AI-generated content if it is not clearly watermarked,” he said.
Legal Consequences
The legality of posting AI-generated images using likenesses of celebrities is also unclear.
“A lot depends on how the content is presented,” said Brent Mittelstadt, research director at the Oxford Internet Institute. “If it's clearly presented as a joke or satire, then you have a pretty strong legal footing.”
“If it's something that is clearly satire to any reasonable person, then you can essentially avoid any legal obligations that way, at least if there are issues of copyright or rights of publicity or whatever,” he said.
In the case of the Swift-related content, Mittelstadt said it's unclear whether the images are being presented as memes or satire, “but based on the content itself, it's clearly not fake to a reasonable person.”
Representatives for Trump did not respond to BI's request for comment.
