eWEEK content and product recommendations are editorially independent. We may make money when you click on links to our partners. Learn More.
Singer Katy Perry shared an AI-generated image of herself attending the Met Gala, referencing a similar image that caught her mother unawares last year. The high-profile share of the photorealistic image highlights how convincing deepfake images can be and how easy it is to potentially create inflammatory pictures.
AI invented a Met Gala dress, hair, and makeup styling for Perry
The picture Perry shared shows the star wearing an outfit matching the 2025 Met Gala theme, “Superfine: Tailoring Black Fashion.” The annual Met Gala raises money for the Metropolitan Museum of Art’s Costume Institute and serves as an avant-garde fashion show populated by the rich and famous. The AI-generated costume combines a futuristic-looking dress with a formal, pinstriped suit.
“couldn’t make it to the MET, I’m on The Lifetimes Tour (see you in Houston tomorrow IRL‼️),” Perry wrote on Instagram, referring to her ongoing concert tour promoting her newest album, “143.” “P.s. this year I was actually with my mom so she’s safe from the bots but I’m praying for the rest of y’all.”
In 2024, Perry made headlines with an Instagram post showing another AI-generated image of herself at the Met Gala, along with a text from her mother. The text included “Didn’t know you went to the Met. What a gorgeous gown.” Perry responded with “lol mom the AI got you too, BEWARE!”
Perry’s post can be seen as a lighthearted note to her followers to be aware of more possible dupes. She did not share who had created the deepfake or what generative AI model had been used.
AI-generated images of individuals are not allowed by OpenAI – unless they’re public figures
One of the most well-known AI companies, OpenAI, loosened its image generation content moderation policies in March 2025 to allow ChatGPT to generate images of public figures. Previously, OpenAI’s image generation tools would discourage photorealistic images of real people in an effort to cut down on misleading or otherwise harmful political imagery.
The same update allowed the model to replicate the styles of living artists, leading to the flood of Studio Ghibli-inspired images earlier this year. Artists protested the automated replication of such distinct styles, and then OpenAI put a stop to the trend, citing “important questions and concerns within the creative community.”
Currently, OpenAI’s usage policies state: “You may not edit images or videos that depict any real individual without their explicit consent,” as well as “4o image generation is capable in many instances of generating depictions of a public figure based on a text prompt” and providing instructions for public figures who wish for their image not to be generated. Of course, users tend to find ways around individual companies’ guidelines.
Read eWeek’s coverage about actress Scarlett Johansson calling out the use of her likeness in a deepfake video. And, that judges of the 2026 Academy Awards will not take into account whether generative AI tools were used in the making of a movie.