Some of the biggest stars in music, film and fashion walked the Met Gala’s red (or, in this case, green) carpet this year.

One star who wasn’t there? Katy Perry. And yet, after the event got underway, the pop star says she received a confusing text from her mom.

Keith Hudson, from left, Katy Perry, Mary Hudson arrive at the 35th Annual Colleagues Spring Luncheon and Oscar de la Renta Fashion Show on April 25, at the Beverly Wilshire Hotel in Beverly Hills, Calif. Perry’s mom, Mary Hudson, was also duped by fake AI pictures of Perry at the Meta Gala.  Jordan Strauss/Invision via Associated Press

“Didn’t know you went to the Met,” Perry’s mom said, according to a screenshot of the message posted by the star on Instagram. “What a gorgeous gown, you look like the Rose Parade, you are your own float lol.” In the screenshot, the text included a photo of what appeared to show Perry in a dramatic flower-trimmed gown, surrounded by photographers.

Except the photo wasn’t real: It was one of several fake images appearing to show stars at the Met that were probably generated by artificial intelligence and spread widely on social media – a scenario that has prompted past warnings from experts about the risks of using technology to manipulate imagery.

Perry’s mom, Mary Hudson, was not the only one who apparently fell for the trick. One post that shared the fake image of Perry was viewed 16.6 million times on X, prompting users to add a context label using the platform’s “community notes” function – and highlighting a major challenge of AI-generated images: It’s quick and easy for fake or manipulated photos to spread online, and much slower to fact-check and add warnings to them.

“lol mom the AI got you too, BEWARE!” Perry wrote to her mother, according to her screenshot on Instagram.


Nor was Perry the only celebrity with fake images circulating after the Met Gala: A fake image of Rihanna attending, dressed in a tree-shaped gown covered in vines, flowers and birds, also made the rounds late Monday and early Tuesday – even though the singer and entrepreneur wasn’t at the event, reportedly because she was sick with the flu. Manipulated images of Selena Gomez were also shared online, though the actress also didn’t attend.

Perry wrote on Instagram she “couldn’t make it to the MET” because she “had to work.” In her post, she included two of the widely shared AI-generated images that falsely appeared to show her at this year’s event. Representatives of the pop star did not immediately respond early Wednesday to a request for comment about the images.

Fake or manipulated photos, videos and even audio are an increasingly common phenomenon, particularly around big events that capture the internet’s attention.

The images can appear harmless: Last year, for example, a fake photo of Pope Francis in a puffy white coat labeled as being from the brand Balenciaga didn’t do much more than prompt many to praise the religious leader’s apparently modern style – but the ease with which people fell for it highlighted the risks of content that appears just real enough to confuse, mislead or convince us of something that didn’t actually take place.

Experts say the proliferation of false or distorted imagery can promote misinformation, and in more sinister cases, allow for scams, blackmail, identity theft and the manipulation of current events.

Earlier this year, the Justice Department announced a new effort to combat the use of AI for criminal activity, as well as to create new rules around its own use of AI. Deputy Attorney General Lisa O. Monaco said in a speech that AI was both a threat and a potential tool.


“It has the potential to be an indispensable tool to help identify, disrupt, and deter criminals, terrorists, and hostile nation-states from doing us harm,” she said. “Yet for all the promise it offers, AI is also accelerating risks to our collective security.” She said AI can be biased and promote discrimination, as well as create “harmful content” such as child porn, make it easier for criminals to commit crimes, and provide authoritarian governments with the tools to manipulate and repress their populations.

As The Washington Post previously reported, artificial intelligence is already sowing confusion in the 2024 presidential election – and not just because of false or manipulated audio, video and images. In one incident, former president Donald Trump dismissed an ad on Fox News featuring video of his well-documented public gaffes – including his struggle to pronounce the word “anonymous” in Montana and his visit to the California town of “Pleasure,” a.k.a. Paradise, both in 2018 – claiming the footage was generated by AI.

The Lincoln Project, a political action committee formed by moderate Republicans to oppose Trump, was quick to deny the claim; its ad featured incidents that were widely covered at the time and witnessed in real life by many independent observers.

Still, AI creates a “liar’s dividend,” Hany Farid, a professor at the University of California at Berkeley who studies digital propaganda and misinformation, previously told The Post. “When you actually do catch a police officer or politician saying something awful, they have plausible deniability” in the age of AI.

Some low-quality AI-generated images are easy to spot – for example, by examining elements that seem off, such as a person with an extra finger, or an artificial sheen on people’s skins, or by looking at the context. One of the fake photos of Perry at the Met Gala appeared to feature the carpet from the 2018 gala, not this year’s distinctive greenish carpet.

But Perry’s comment that her own mother fell for the fake image highlights just how much the technology has improved, and how easy it is to be taken in.

Only subscribers are eligible to post comments. Please subscribe or login first for digital access. Here’s why.

Use the form below to reset your password. When you've submitted your account email, we will send an email with a reset code.

filed under: