Clinicians who treat eating disorders say patients are increasingly turning to general-purpose AI chatbots for diet and exercise advice, often receiving guidance that directly contradicts their treatment plans and, in some cases, endangers their recovery.
Health & AI · Eating-Disorder Care
When a Chatbot Becomes a Secret Coach
Clinicians warn that patients are asking general-purpose AI for diet and fitness tips that contradict their treatment plans — advice that is harmless for a healthy person but dangerous for someone in recovery.
AI-related inquiries to an eating-disorder helpline
A roughly 15-fold rise in helpline contacts tied to AI use in about a year.
2023
A purpose-built support chatbot, Tessa, was pulled offline after giving dieting and weight-loss advice.
24/7
Constant availability is the draw — but the model can't tell a bodybuilder from a malnourished patient.
The Core Risk
Context-blind advice, delivered at scale, can quietly work against months of careful treatment.
• The system has no understanding of the patient's clinical situation.
• Overly agreeable models may validate harmful requests.
• These are general-purpose assistants never designed for clinical use.
The Concern
A safety gap in agreeable systems that may endorse harmful goals, with little scientific evidence behind therapy-style tools and failures to spot signs of illness in younger users.
The Upside
AI can expand access where care is limited or costly — and some patients who first consult a chatbot go on to contact a human helpline, acting as an on-ramp to real support.
Open challenge: Guardrails tailored to vulnerable users remain unresolved — developers are urged to learn from the failures of social media rather than repeat them.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…