Why AI food looks like that
The Verge Robert Hart ● Covered by 2 sources
Opinion — commentary, not a factual news event.
Restaurants and brands are using AI to make food images, and the results are weirdly awful. Think donut shrimp, holey burritos, and ice cream that looks like bricks.
Based on reporting by The Verge, Robert Hart — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Restaurants, cafes, and brands are feeding their menus to AI image generators, and the output is a parade of edible-looking mistakes. Instead of a glossy burger or a neat bowl of noodles, the machines keep producing things that look slightly cursed and very hard to eat: donut shrimp, Reubens from the deep, wormlike noodles, noodle-like pastries, and stringy chicken.
The weirdness doesn’t stop at shape. Some of these images drift into full-on material science. There’s construction material pretending to be ice cream, ice cream pretending to be brains, and food that seems to have picked up masonry on the way to the table.
Then there’s the burger. The source doesn’t even try to civilize it, which feels fair. And the burrito gets described as a trypophobic mess, packed with grubs and holes, the kind of image that makes appetite quietly leave the room.
What makes the whole thing stand out is how quickly AI food art slides from imperfect to genuinely grotesque. These aren’t just bad photos. They’re uncanny little accidents that make you wonder whether the prompt was food, decor, or a dare.
My take — AI-written commentary, not fact-checked reporting
This is the logical end of lazy AI marketing: if you won’t hire a food photographer, you get haunted lunch. Brands keep reaching for cheap image generation, then act surprised when the result looks like a catering order from the underworld. The bigger joke is that real food already photographs well — AI just keeps proving that taste is the one thing it can’t fake.
Read more about this at: The Verge
