As the capabilities of artificial intelligence evolve at a rapid pace, legislators in Arizona are working to keep so-called “deepfakes” created with AI from influencing elections and destroying reputations. The state House of Representatives on May 15 approved three bills aimed at punishing those who use AI to create false videos and voice messages meant to embarrass people or extort money from them, as well as to ensure AI-assisted efforts to defame candidates for public office are publicly exposed. …
Whether you love or hate celebrity culture, the Met Gala is an event. Those less jaded among us get to see all of the biggest stars take their boldest fashion risks of the year; unlike an award show, it’s an event that encourages avant-garde extravagance. And if you find the whole thing vapid, then you can laugh along at how Ed Sheeran’s outfit looks like something Troy Bolton wore in High School Musical 3.
But this year, the Met Gala was a test.
“Katy Perry. That’s it,” one X post read. In the photo, Katy Perry appears to be wearing a massive gown decorated with three-dimensional floral appliques. As the pearlescent gown drapes down to the ground, its long train fades into realistic-looking moss, which cascades across the beige and red Met Gala carpet.
If you look at the image for longer than a passing glance, you could feasibly believe that the “Teenage Dream” singer arrived at fashion’s biggest night wearing this whimsical, woodland-like gown. But in all of the other photos of the Met Gala, the carpet was white and green as part of the “Garden of Time” theme; in this photo, it’s beige and red. Why does Katy appear to be walking a different carpet?
That’s the tell-tale sign that this viral image of Perry is fake. And yet it already has more than 10 million views on X and over 300,000 likes.
Minutes later, another user on X posted another image of Perry. Her lips are slightly parted like she’s surprised by the paparazzi, and she’s wearing a bronze corset that looks like a key to a garden. Unlike the first image, this one actually has the correct color palette and scenery for this year’s gala, but something still feels off. Her floral skirt seems like it was cut and paste onto her body, and the light is hitting her corset in an unnatural way.
So, how do you know it’s fake, as opposed to a strange photo? No fashion magazine has reported on Katy Perry’s look tonight — it doesn’t look like she’s actually attending this year. Meanwhile, Perry cryptically liked both viral tweets, but did not comment on the deceptive posts.
Every year at the Met Gala, Rihanna is among the best dressed. In the days leading up to the event, she promised fans that she would actually get to the event in time for dinner — usually, she’s very fashionably late. She even dyed her hair pink for the occasion.
On the earlier end of the red carpet, an image surfaced of Rihanna wearing a dramatically regal, garden-themed dress. The shoulders balloon into a sculptural halo, embroidered with birds, vines and flowers. But again, despite getting 2.6 million views, the image is not real. Rihanna dyed her hair pink, remember? Like Perry, Rihanna is nowhere in sight. People Magazine reported that Rihanna had to skip Monday evening’s festivities after coming down with the flu.
The consequences of a fake Rihanna dress are pretty minor. But based on Rihanna’s history of nailing the Met Gala theme, would she really take the easy way out and wear a floral gown to the “Garden of Time” celebration? Like an AI deepfake Drake song, these synthetic looks were a bit too on the nose, lacking the creativity that makes the Met Gala unique. Could an AI come up with Cynthia Erivo’s brilliant suit or Lana Del Rey’s creepy-cool woodsy look?
Zendaya, an early front-runner for best dressed, showed up relatively in a blue-green outfit that makes her look like a super chic fairytale villain. Hours later, when an image cropped up of Zendaya wearing a black leather gown and floral headpiece on the carpet, I was braced to believe that it was yet another fake photo. But the truth is stranger: After five years absent from the gala, Zendaya actually walked the carpet twice in two different outfits. Go figure.
Still, each new image of a celebrity served as a call to check the patterns on the carpet, the flowers along the stair railings, and whether or not the paparazzi in the background look a little funky. Usually, the Met Gala is an opportunity to gab about famous people’s outlandish outfits as a brief distraction from our unglamorous Monday nights. But in the age of widespread generative AI tools, celebrity culture serves as a constant reminder that we can’t believe everything we see online.
Hanks shared a screenshot of the ad to Instagram with the caption, “Beware!! There’s a video out there promoting some dental plan with an AI version of me. I have nothing to do with it.”
Story continues below advertisement
The AI-generated Hanks appears to have been aged down in the advertisement.
Hanks did not reveal the company name or where the advertisement was being shown, and his representatives have declined comment.
Gayle King issues a similar warning
Only one day after Hanks’ warning, Gayle King also shared a message warning her followers about a fraudulent video advertisement using her likeness. The CBS Mornings host said malicious advertisers used AI to manipulate a legitimate video of King promoting her radio show in August.
“People keep sending me this video and asking about this product and I have NOTHING to do with this company,” King wrote. “They’ve manipulated my voice and video to make it seem like I’m promoting it.”
Story continues below advertisement
King posted the AI weight loss advertisement to Instagram, with the words “Fake Video” stamped overtop. She also included footage from her original radio show promotion to prove that the advertisement had been doctored.
“I’ve never heard of this product or used it! Please don’t be fooled by these AI videos,” she concluded.
King did not specify where the advertisement was being shown.
Meta, the parent company of Instagram and Facebook, told the New York Times in an email that it is “against our policies to run ads that use public figures in a deceptive nature in order to try to scam people out of money.”
Story continues below advertisement
“We have put substantial resources towards tackling these kinds of ads and have improved our enforcement significantly, including suspending and deleting accounts, pages and ads that violate our policies,” the statement continued.
‘Disturbing’ AI recreations
Even after death, celebrities are not immune to AI manipulation.
Zelda Williams, the daughter of iconic comedian and actor Robin Williams, on Sunday said her father’s image and voice had been replicated by AI and shared online.
In a statement posted to her Instagram story, Zelda called AI-generated content featuring her father’s likeness “disturbing.”
“I am not an impartial voice in SAG’s fight against AI,” Zelda wrote. “I’ve witnessed for YEARS how many people want to train these models to create/recreate actors who cannot consent, like Dad. This isn’t theoretical, it is very very real.
“I’ve already heard AI used to get his ‘voice’ to say whatever people want and while I find it personally disturbing, the ramifications go far beyond my own feelings,” Zelda continued. “Living actors deserve a chance to create characters with their choices, to voice cartoons, to put their HUMAN effort and time into the pursuit of performance.”
Zelda called AI-generated versions of actors “at their worst, a horrendous Frankenstein monster, cobbled together from the worst bits of everything this industry is, instead of what it should stand for.”
This is far from the first time actors like Hanks have expressed worry about the use of AI in the entertainment industry.
Earlier this year, Hanks voiced his concern about AI and internet deepfakes when he appeared as a guest on The Adam Buxton Podcast. Hanks said he first began to worry after starring in the 2004 Christmas film The Polar Express, which saw Hanks and a cast of others animated using motion-capture technology.
“We saw that there was going to be this ability in order to take zeros and ones inside a computer and turn it into a face and a character,” Hanks said in the interview. “Now, that has only grown a billion-fold since then and we see it everywhere.”
On The Adam Buxton Podcast, Hanks said there were “discussions going on in all of the guilds, all of the agencies and all of the legal firms in order to come up with the legal ramifications of my face and my voice — and everybody else’s — being our intellectual property.”
Hanks said the technology could allow studios to perpetually create movies starring AI versions of him at various, younger ages. He said he could be “hit by a bus tomorrow” and still have his likeness star in future performances.
“Outside of the understanding that it’s being done by AI or deepfake, there’ll be nothing to tell you that it’s not me and me alone — and it’s going to have some degree of lifelike quality,” he said. “That’s certainly an artistic challenge, but it’s also a legal one.”
The WGA strike was declared over this month after board members approved a contract agreement with studios. The agreement prohibits studios from using AI to write or rewrite material.
Hollywood writers, studios reach tentative deal to end strike
SAG-AFTRA members are still on strike.
Story continues below advertisement
In Canada, the Liberal government last year introduced legislation that proposed new rules for the use of AI. The proposal came as part of a federal privacy bill to give Canadians more control over how their personal data is used by commercial entities.
Government introduces new privacy bill to give Canadians more control over online data
Numerous civil society organizations, experts and academics have called on the government to amend its proposal so that AI is considered separately. Advocates have argued the existing AI section of the bill fails to protect the rights and freedoms of people from the risks of AI.
