Fake photos, but make it fashion. Why the Met Gala pics are just the beginning of AI deception

Actor Jared Leto carrying around his own head as an accessory? Real. Rapper Lil Nas X, painted head to toe in silver, his body encrusted with pearls and crystals, wearing only a metallic Dior thong? It happened. Actor and singer Billy Porter, wearing a catsuit, carried into the event by six shirtless men in gold pants? Yes. 

If there’s any event where it might be difficult to discern reality from fantasy, it’s the Met Gala, where Grimes once brandished a sword and Lady Gaga once stripped through four different outfits until she was wearing only a black lingerie set, go-go boots, and pulling a pink wagon behind her on the carpet.

But this year, people weren’t tripped up by the fashion choices (which were relatively tame, naked dresses aside). Instead, they were confused about which celebrities were actually there, thanks to AI-generated images during fashion’s biggest night.

And while the AI photos swirling online of celebrities like Katy Perry and Rihanna might seem harmless, experts note that each instance of people being misled by generative AI underlines growing concerns around the misuse of this technology.

It’s particularly concerning regarding disinformation and the potential to carry out scams, identity theft or propaganda, and even election manipulation, they said.

“It used to be that seeing is believing, and now seeing is not believing,” Cayce Myers, a professor and director of graduate studies at Virginia Tech’s School of Communication, told the Associated Press.

“[If] even a mother can be fooled into thinking that the image is real, that shows you the level of sophistication that this technology now has.”

Katy Perry wasn’t at the Met Gala

AI-generated images depicting a handful of big names, including Perry and Rihanna, at the Metropolitan Museum of Art’s annual fundraiser quickly spread online Monday and early Tuesday. Perry and Rihanna didn’t attend the gala, and it’s unclear where exactly the photos of them originated from.

Perry re-posted two of the images to her Instagram Monday night, writing that she “couldn’t make it to the Met, had to work.” She also included a video of herself in the studio, and a screenshot of a text from her mom complimenting her outfit.

“lol mom the AI got to you too, BEWARE!” Perry responded in the text screenshot.

By Wednesday morning, the post contained a warning from fact-checker PolitiFact that the Met photo was AI-generated.

Fake photos of other celebrities from the gala circulated online, as well, including one of Rihanna in a white gown, and another of singer Jason Derulo falling down a flight of stairs. On X, formerly Twitter, a warning was added to the photos of Rihanna, saying they’re AI.

The fact-checking website VerifyThis confirmed Derulo did not fall down the stairs — the photo shared isn’t of him, and was in fact taken at the Cannes Film Festival in 2011.

But first, the pictures and others like them were shared and liked across social media platforms. An influencer on TikTok even gave a glowing review of a fake outfit on Kim Kardashian. (As Forbes pointed out, Kardashian was at the event, but wearing something else.) The video has 9.6 million views.

In a follow-up video, the influencer reacted to Perry’s fake image, saying “the fact that her own mother, her own blood, thought this was Katy Perry … the entire internet was fooled.”

The problematic rise of photo manipulation and AI

This is far from the first time we’ve seen generative AI, a branch of AI that can create something new, used to create phony content. Image, video and audio deepfakes of prominent figures, from Pope Francis to Taylor Swift, have gained loads of traction online before.

It’s also far from the first photo manipulation that had people questioning what’s real, and has media outlets looking at photos with renewed scrutiny. 

In March, for instance, two images released by Kensington Palace were found to have been digitally altered — the now-infamous photo of Catherine, the Princess of Wales, surrounded by her smiling children; and a 2022 image of Queen Elizabeth with her grandchildren and great-grandchildren.

But AI has taken these concerns to the next level. 

“The implications here go far beyond the safety of the individual — and really does touch on things like the safety of the nation, the safety of [our] whole society,” David Broniatowski, an associate professor at George Washington University and lead principal investigator of the Institute for Trustworthy AI in Law & Society at the school, told the Associated Press. 

Earlier this year, sexually explicit and abusive fake images of Swift, for example, began circulating online. Research also shows that explicit AI-generated material overwhelmingly harms women and children — including cases of AI-generated nudes circulating through high schools.

In March, the Center for Countering Digital Hate (CCDH), a U.K. non-profit that monitors online hate speech, released a report showing how AI image generators can threaten election integrity and democracy.

The centre used generative AI tools to create images of U.S. President Joe Biden lying in a hospital bed and election workers smashing voting machines, raising worries about potential falsehoods ahead of the U.S. presidential election in November.

“The potential for such AI-generated images to serve as ‘photo evidence’ could exacerbate the spread of false claims, posing a significant challenge to preserving the integrity of elections,” CCDH researchers said in the report.

And that’s where the fake Met gala photos are making some people nervous.

“The AI generated fake photos from the Met Gala are a low-stakes prelude for what’s going to happen between now and the elections,” a user wrote on X.

“Watching everyone get fooled in real-time by the AI Rihanna Met Gala look should make us all quake in fear about the upcoming election coverage. I’m so tired already,” wrote another.