New York Met Gala | Faked photos of stars are invited to the gala

(Paris) American singer Katy Perry deceived many Internet users by publishing a photo of herself doctored by artificial intelligence (AI) on the steps of the Met Gala, marked this year by a series of deepfakes of celebrities.


“I couldn’t go to the MET, I had to work,” wrote the star on Instagram (207 million subscribers) in a message accompanying a photo where she can be seen in an oversized floral dress, frozen gaze, on what appears to be the Metropolitan Museum gala carpet.

A second photo shows her wearing a metallic bustier and a low-cut imitation floral petticoat, in an outfit that looks like an Amazon warrior.

Tuesday morning, the singer’s publication, geolocated in the metaverse, had more than 1.2 million views. likes on Instagram. These AI montages (or deepfakes) published by the interpreter of Firework were then found on other social networks like X, where many Internet users believed them to be authentic.

Among them, the artist’s own mother, as evidenced by a screenshot shared by Katy Perry: “I didn’t know you went to the Met” we can read in this exchange, to which she responds “Lol Mom, you too were fooled by the AI, BE CAREFUL! »

Others deepfakes stars who were not present at the gala, such as Rihanna and Selena Gomez, spread across social networks.

On the most popular publications containing these images, messages now warn Internet users on X that these are photos created with AI.

If, like Katy Perry, certain personalities have seized these tools for their communication, the rise of generative AI has given new impetus to the phenomenon of deepfakesin particular manipulated and sexualized images that feature women, for the purposes of intimidation or harassment.

In January, a fake pornographic image of superstar Taylor Swift was viewed 47 million times on X before being deleted by the social network around fifteen hours later.

In February, twenty digital giants, including Meta (Instagram, Facebook), TikTok and X, committed to deploying technologies to more effectively identify and report misleading content generated by AI.


source site-52