Taylor Swift’s endorsement of Kamala Harris this week should teach hard truths to two men: JD Vance should never have spoken ill of “childless cat ladies.” And Donald Trump should never have messed around with artificial intelligence.
After the presidential debate Tuesday night, Swift posted a photo of herself that amounted to a subtweet of Vance — looking into the camera head on, holding her fluffy, blue-eyed cat, a ragdoll named Benjamin Button. The caption was a letter to her fans, explaining why she plans to vote for Harris, citing LGBTQ+ rights, in vitro fertilization, abortion and, incongruously, deepfakes.
The unexpected addition to the list of progressive policy priorities was a response to images that Trump, perhaps jokingly, posted to Truth Social late last month. “Swifties for Trump,” read the signs in one photo. “Taylor wants YOU to vote for Donald Trump,” says the text on another that features a robotically generated Swift posing Uncle Sam-style. Swift, however, does not even want to vote for Donald Trump herself, and she set the record straight.
“Recently I was made aware that AI of ‘me’ falsely endorsing Donald Trump’s presidential run was posted to his site,” Swift wrote Tuesday. “The simplest way to combat misinformation is with the truth.”
That Swift reacted poorly shouldn’t have come as a surprise to the former president. (Trump didn’t apologize, but he did admit the “photos” were fabricated by somebody else.) The pop star’s desire to protect her image, her intellectual property and her image as her intellectual property is lesson 101 in Swiftology.
Still, as much as this episode is what it says about Swift, it says even more about generative artificial intelligence. We worry about robots automating jobs out of existence, or hallucinating wrong answers to high-stakes questions, or producing false news stories on command — and we’re right to. But perhaps the most troubling threat the technology poses is to who we are. The idea that we might lose agency over our faces, our feelings and the fruits of our creativity is just as personal for any of us as it is for Swift.
The AI imbroglio might not alone have prompted Swift to tip her hat to Harris: She endorsed Joe Biden last time around, also on the night of a debate, bringing extra attention to her even as she brought more to politics. The brouhaha over her friendship with Brittany Mahomes, who “liked” (and then later “unliked”) a pro-MAGA Instagram post, might have moved up the timing, too.
Still, given her record on protecting her likeness, it’s improbable that stealing the image of the most famous woman on the planet for personal gain would do nothing to sway her toward issuing a political statement in line with her past endorsements.
Swift has the means to fight back and the inclination. She is notoriously litigious. She countersued a radio host, who sued her for defamation, for sexual assault and battery. She threatened legal action against a college student whose tracking of her private jet got her in trouble with environmental activists angry about carbon dioxide emissions. And those aren’t even her greatest hits. Her infamous feud with Kanye West reached its zenith when he included her as a nude wax figure in bed with other celebrities in his “Famous” music video.
She described that incident as “revenge porn.” This was a stretch — but in a recent run-in with generative AI that predated the “Taylor wants YOU” non-endorsement, her description was closer to the mark. This year, deepfake pornography of the star swirled throughout the internet.
Swift cares as intensely as perhaps anyone about her literal image and her image as an idea: the world’s perception of who she is as an artist and a person. Generative AI, the best copycat we’ve ever seen, tried to steal both from her — by painting her as a Trump supporter, and by just plain painting her.
The rise of this technology presents less of a financial risk, but as much a personal danger to the rest of us. Our own images and our very selves are at more risk than ever. At its most frightening, generative AI allows malicious actors to manipulate our faces and voices to tell damaging lies about us. Yet even these systems’ less menacing ability to hoover up pictures, words and the ideas they convey means they can mimic us and co-opt the creativity that makes us human.
That’s the bad news. The good news is, when misinformation involves someone, or something, pretending to be us, not only do we feel compelled to counter it, but we’re also in a prime position to prove we’re right. All we have to do is say basically what Swift said with her endorsement: That’s not me.
Getting the world to listen may be easier when you’re a pop star. Yet her words should set an example to everyone else who feels their livelihoods or their selves are threatened by AI to — as the singer-songwriter herself would put it — speak now.
