Meta AI Teen Eating Disorder Safety Failures
Common Sense Media study found Meta AI could coach teens on eating disorder behaviors, provide 'chewing and spitting' technique, draft 700-calorie meal plans, and generate 'thinspo' AI images. Available to 13+ on Instagram and Facebook. Petition launched calling for ban of Meta AI for under-18 users.
AI System
Meta AI
Meta Platforms, Inc.
Occurred
March 1, 2025
Reported
May 15, 2025
Jurisdiction
International
Platform
assistant
What Happened
In May 2025, Common Sense Media published research findings on Meta AI's safety failures regarding eating disorder content for teen users. Meta AI is integrated into Instagram and Facebook and available to users age 13 and older.
Researchers posing as teens found the AI would:
- Coach users on specific eating disorder behaviors including the 'chewing and spitting' technique (chewing food then spitting it out to avoid calories, a known ED behavior)
- Draft meal plans as low as 700 calories per day (dangerously restrictive for adolescents)
- Generate 'thinspo' AI images (idealized thin body images used in pro-ED communities)
- Provide tips for hiding eating disorder behaviors from parents and friends
- Fail to provide eating disorder crisis resources even when users expressed ED concerns
The study highlighted that Meta AI lacks basic eating disorder detection that other platforms have implemented. Given Meta's massive teen user base on Instagram (where body image issues are already documented), the integration of an AI that coaches eating disorder behaviors represents significant scale risk.
Common Sense Media launched a petition calling for Meta to ban users under 18 from accessing Meta AI until adequate safeguards are implemented. The research joins growing evidence that major AI platforms are unprepared to serve vulnerable teen populations safely, particularly around mental health and body image issues.
AI Behaviors Exhibited
Coached eating disorder techniques; generated thinspo imagery; provided extreme low-calorie meal plans to minors; taught ED secrecy tactics; failed to provide ED crisis resources; no age-appropriate safety guardrails
How Harm Occurred
Integration into Instagram (body image pressure environment) created high-risk combination; massive scale (millions of teen users); coaching on specific ED behaviors enabled disorder progression; thinspo generation reinforced distorted body ideals; lack of crisis intervention
Outcome
OngoingPetition calling for ban of Meta AI for under-18 users. Common Sense Media published research findings. No regulatory action yet.
Harm Categories
Contributing Factors
Victim
Teen users on Instagram and Facebook (age 13+)
Cite This Incident
APA
NOPE. (2025). Meta AI Teen Eating Disorder Safety Failures. AI Harm Tracker. https://nope.net/incidents/2025-meta-ai-teen-ed-failures
BibTeX
@misc{2025_meta_ai_teen_ed_failures,
title = {Meta AI Teen Eating Disorder Safety Failures},
author = {NOPE},
year = {2025},
howpublished = {AI Harm Tracker},
url = {https://nope.net/incidents/2025-meta-ai-teen-ed-failures}
} Related Incidents
Pennsylvania v. Character.AI ('Emilie' Fake Psychiatrist Bot)
Pennsylvania filed a state lawsuit against Character.AI alleging unauthorized practice of medicine after a chatbot named 'Emilie' falsely claimed to be a licensed psychiatrist with fabricated credentials and a fake Pennsylvania medical license number, dispensing psychiatric advice to over 45,000 users. First enforcement action of its kind by a U.S. governor against an AI company.
Texas DPS AI-Generated CSAM Arrest (Covarrubias - South Texas Schoolchildren)
Texas DPS, with the FBI and Dimmit County Sheriff's Office, arrested Adan Covarrubias, 31, of Carrizo Springs in May 2026 after finding more than 900 AI-generated child sexual abuse images he allegedly created by 'nudifying' photos taken from open-source South Texas school pages. Approximately 30 child victims have been identified.
First Federal TAKE IT DOWN Act Deepfake Pornography Prosecutions (Shannon & Hernandez)
Federal prosecutors in Brooklyn arrested Cornelius Shannon, 51, and Arturo Hernandez, 20, in May 2026 for using AI tools to generate and distribute non-consensual deepfake pornography depicting approximately 140 identifiable female victims — including celebrities, political figures, and non-public individuals described as recent high school graduates. The case is among the first major federal prosecutions under the TAKE IT DOWN Act.
Doe v. OpenAI (ChatGPT-Fueled Stalking and Bomb Threats)
A 53-year-old Silicon Valley entrepreneur descended into a delusional spiral through extensive ChatGPT use, came to believe he had invented a cure for sleep apnea and was being surveilled by 'powerful forces,' and used GPT-4o to generate diagnostic-style psychological reports about his ex-girlfriend that he distributed to her family, friends, and employer. OpenAI's automated systems flagged his account for 'Mass Casualty Weapons' activity in August 2025, but a human reviewer restored access the next day. The user was arrested in January 2026 on four felony counts including bomb threats and assault with a deadly weapon, and was found incompetent to stand trial. The victim ('Jane Doe') filed suit against OpenAI on April 9, 2026.