Eating disorder therapists say patients are increasingly using AI chatbots, undermining their advice; AI also appears to be directing more patients to helplines
Context & Ripple Effects
Related coverage has traced a progression from AI chatbots being used as low-cost, always-available supplements or substitutes for mental-health services to clinicians reporting that these interactions can deepen negative feelings. Professional concerns have also reached the FTC, with the American Psychological Association warning about chatbots presenting themselves as therapists.
This report narrows that broader concern to eating-disorder care: chatbot use is becoming part of the treatment relationship itself, sometimes conflicting with clinicians’ guidance while also apparently steering some users toward crisis resources.
First-order effects
- Eating-disorder therapists must account for chatbot conversations as an active influence on patients’ decisions and adherence to treatment advice.
- More chatbot referrals to helplines could route some people toward human support, but conflicting guidance can complicate clinical care for patients already in treatment.
Second-order effects
- Therapists, providers, and chatbot makers face stronger pressure to define when an AI tool should defer to clinicians or direct users to crisis and specialist services.
- The gap between chatbots’ convenience and the limits of unsupervised mental-health support becomes more consequential for vulnerable user groups, raising the stakes around product safeguards and disclosure.
Third-order effects
- If clinicians continue to encounter AI as a competing source of mental-health guidance, care may shift toward models that treat chatbot use as something to be monitored and integrated rather than ignored.
- The pattern strengthens the case for clearer accountability over mental-health-oriented chatbot behavior, particularly where products can influence high-risk users without a human therapeutic relationship.
The trend: Consumer AI is moving from a general-purpose support tool into sensitive care pathways, forcing providers and policymakers to confront its role alongside—not outside—professional services.