AI bodies worry local eating disorder experts
Add Axios as your preferred source to
see more of our stories on Google.
Illustration: Lindsey Bailey/Axios
Kids are measuring themselves against AI-generated bodies that no one actually has, local mental health experts tell Axios, and it's showing up in eating disorder treatment.
The big picture: The images hurt body satisfaction even when viewers know they aren't real, research shows, and hospitals already treat hundreds of kids each year for eating disorders.
Driving the news: Some of the fake accounts pose as fitness influencers or sell wellness products, while others post AI-made before-and-after body transformations. Axios reviewed dozens of such posts, and many weren't labeled as AI.
By the numbers: Around 240 kids under age 18 in Jackson, Clay, Platte and Cass counties were hospitalized with an eating disorder diagnosis in 2023, up from around 170 in 2019, according to Missouri Department of Health data obtained by Axios. Jackson County accounted for more than half each year.
- The state's 2024 and 2025 counts are still provisional, so they can't be compared yet, but department officials say early data shows that number climbing.
- The Kansas Department of Health and Environment didn't respond to Axios' request for data.
What they're saying: "When people compare themselves with those images, they may be comparing themselves against something that is edited, selectively presented, or may not even depict a real person," Kelsie Forbush, director of KU's Center for the Advancement of Research on Eating Behaviors, tells Axios.
- Her lab has already found a link between eating disorder behavior and eating disorder-related Instagram content, which is why she says parents should keep an eye on what their kids are seeing.
- She suggests that parents talk with their kids about the images. "It's a matter of helping their child align their values and interests with how they're spending their time (aka, getting them off their phone)," she says.
- Missouri Eating Disorders Council chair Annie Seal told Axios that the council is watching AI closely with "particular concern" about content pushing unrealistic standards for weight, appearance and dieting.
State of play: Don't count on the platforms to flag it. Meta lets users opt in to labeling AI content and may flag some itself, and TikTok "encourages" creators to label AI-generated or heavily edited posts.
- Neither rule is foolproof, and likely AI content still shows up unlabeled.
- Missouri's closest attempt at a fix was a House bill to keep kids under 16 off social media without a parent's consent. It passed 145-3 in April, then stalled in a Senate committee after tech industry groups opposed it.
If you need help, the National Alliance for Eating Disorders runs a free helpline at 866-662-1235.