Can AI Support Anorexia Treatment?
Warning: This post mentions suicidal ideation and attempt. If you are thinking about suicide, or worried about someone who might be at risk, please call or text 988 Canada for support. If you think someone is in imminent danger, call 911.
In the first few posts of this blog series on AI and anorexia I talked about what anorexia is, the risks, and how it is treated.
In my previous post about treatments, I describe some of therapies used, I mentioned that some gaps exist in popular models.
This post describes some of those gaps, while also bringing forward interesting advances in AI technology. Some of these technologies have are already being explored for medical and mental health purposes, but there may be emerging opportunities for adapting AI to anorexia treatment.
Limitation 1: Access to Treatment
Barriers include cost, geographical location, wait times, lack of mental health literacy, stigma, and more.
Food for Thought
GenAI might reduce barriers such as stigma, geographical limitations, and financial constraints by offering private, accessible, and scalable tools like chatbots and virtual platforms. AI-based tools may bridge gaps during long wait times for inpatient treatment, providing interim support to prevent symptom deterioration.
Limitation 2: Personalization of care
Manualized and heavily structured programs might lack personalization, and one-size-fits-all approaches are not always enough.
Food for Thought
GenAI can analyze patient data to create individualized treatment plans, dynamically adapting to the unique needs of each patient. Tools like wearable technology and machine learning (ML) algorithms can monitor health metrics (e.g., heart rate, sleep patterns) to inform treatment adjustments.
Limitation 3: CBT is fine but...
There are lots of ways that CBT could be improved, and more data needs to be collected to find out how effective it actually is. Some people receiving inpatient care have reported that treatment can feel cold and dehumanizing at times.
Food for thought
AI-powered chatbots might provide CBT-based interventions, emotional regulation tools, and mindfulness exercises to complement human-led therapy. Maybe having an AI companion would help create a sense of personalization.
Limitation 4: Exposure therapy can be hard to deliver
Food exposure works. Increasing the variety of foods a person with anorexia eats has been shown to improve symptoms and might also create empowerment, choice, and agency.
Food for thought
Virtual reality (VR) and GenAI can simulate feared food or body-related scenarios, helping patients confront anxieties in a controlled environment. VR exposure has been shown to elicit similar emotional responses to real-life exposure. This might be a more relaxed or approachable method for some. For example, those with intense food and weight gain fears might benefit from an approach like Gen-AI assisted VR exposure, which seems to straddle the line between real and imagined exposure.
Limitation 5: Externalization is helpful, but too abstract for some
Externalization and creating psychological 'distance' between the self and anorexia has been shown to be effective, but some people find it too abstract.
Food for Thought
GenAI-based AVATAR therapy was created to help people with distressing inner voices, or auditory hallucinations, create digital manifestations of the voice. This aims to give them an opportunity to improve their autonomy and power over these voices. This technology is being trialed for anorexia, but more research is needed.
GenAI-based virtual space creation might also help people externalize their thoughts and feelings using digital objects. Similar to sandplay therapy, virtual 'worlds' might be a ways for people to describe their inner worlds with more ease and access. These types of AI technologies may help patients externalize their eating disorder, fostering identity exploration and empowerment.
Limitation 6: Ways to prioritize gender identity
Including one's gender identity in therapy is be an important aspect to inclusive, socially just, and individualized anorexia treatment.
Food for Thought
Virtual spaces can be empowering for people who identify as 2SLGBTQ+. While there are risks with exploring one's gender and sexuality online, some people report that digital spaces can feel safer and affirming. AI tools can provide safe, gender-affirming, and inclusive spaces for marginalized groups, such as sexual and gender minorities, who face higher risks of anorexia. GenAI-based photo augmentation apps, for example, might provide opportunities for people to be curious and creative when exploring their gender.
An important word of caution
While these platforms may offer community, support, and safer spaces for expression, they can also expose users to harassment, discrimination, and cyberbullying. Protections are needed to help mitigate risk, especially to groups of people who are particularly vulnerable such as children, youth, and people of intersectional marginalized identities. For example, the same photo apps that could help folks explore gender may offer inroads for predators to use these tools for nefarious purposes.
Limitation 7: Weight stigma
People in 'normal' or larger bodies might face stigma or be overlooked by systems, despite their suffering. People with atypical anorexia have many of the same physiological risks as those who are underweight, and the mental distress is just as debilitating.
Food for thought
GenAI may help mitigate biases in diagnosis and treatment by focusing on pathology rather than weight or body size. When weight / BMI is removed from intake and assessment measures, how many more people would find themselves feeling validated, heard, and deserving of care?
Limitation 8: Suicide risk and prevention
People with anorexia are more likely to think about or act on suicide than people without the disorder. The risk of death by suicide in those with anorexia is estimated to be 18 times higher compared to age- and gender-matched individuals without the disorder.
Food for thought
Natural Language Processing (NLP) algorithms may be able to analyze language patterns in counselling sessions, on digital / social media platforms, and in other data sources to identify suicide risk.
While all counsellors in BC are required to monitor and mitigate suicide risk, new tech might offer an extra layer of protection and potential for early intervention.
An important word of caution
Suicide and AI are a complex topic. There have been news articles covering the way that AI seems to influence people who have suicidal ideation, and some accounts of AI being partly responsible for teens' death by suicide. For example this 14 year old and this 16 year old. The dichotomy of AI is that it can be both helpful and harmful, and it is up to humans, companies, and policy-makers to ensure that safeguards are in place. More on this topic in my next post on risks and ethics.
Share Your Thoughts
What are your thoughts on the current state of anorexia nervosa treatment? How can we work together to reduce stigma and improve access to care? Share your thoughts in the comments below!
Sources
American Psychiatric Association. (2022). Feeding and eating disorders. In Diagnostic and Statistical Manual of Mental Disorders (5th ed., text rev.).BC Association of Clinical Counsellors. (2023, November). Code of Ethical Conduct. https://bcacc.ca/wp-content/uploads/2023/07/BCACC-COEC-Effective-Nov.-1-2023.pdf
Clus, D., Larsen, M. E., Lemey, C., & Berrouiguet, S. (2018). The use of virtual reality in patients with eating disorders: Systematic review. Journal of Medical Internet Research, 20(4), e157. https://doi.org/10.2196/jmir.7898
Duggan, H. C., Hardy, G., & Waller, G. (2025). Cognitive-behavioural therapy (CBT) for outpatients with anorexia nervosa: A systematic review and meta-analysis of clinical effectiveness. Cognitive Behaviour Therapy, 0(0), 1–46. https://doi.org/10.1080/16506073.2025.2465745 Gan, J. K. E., Wu, V. X., Chow, G., Chan, J. K. Y., & Klainin-Yobas, P. (2022). Effectiveness of non-pharmacological interventions on individuals with anorexia nervosa: A systematic review and meta-analysis. Patient Education and Counseling, 105(1), 44–55. https://doi.org/10.1016/j.pec.2021.05.031
Gaudiani, J. L. (2018). Sick Enough: A Guide to the Medical Complications of Eating Disorders (1st edition). Routledge.
Greca, A. D., Amaro, I., Barra, P., Rosapepe, E., & Tortora, G. (2024). Enhancing therapeutic engagement in mental health through virtual reality and generative AI: A co-creation approach to trust building. 2024 IEEE International Conference on Bioinformatics and Biomedicine (BIBM), 6805–6811. https://doi.org/10.1109/BIBM62325.2024.10822177
MacDonald, D. E., McFarlane, T., & Trottier, K. (2020). Chapter 22: Eating disorders. In J. E. Maddux & B. A. Winstead (Eds.), Psychopathology: Foundations for a Contemporary Understanding (5th Edition, pp. 495–521). Routledge.
Magrini, M., Curzio, O., Tampucci, M., Donzelli, G., Cori, L., Imiotti, M. C., Maestro, S., & Moroni, D. (2022). Anorexia nervosa, body image perception and virtual reality therapeutic applications: State of the art and operational proposal. International Journal of Environmental Research and Public Health, 19(5), 2533. https://doi.org/10.3390/ijerph19052533
Riva, G., Malighetti, C., & Serino, S. (2021). Virtual reality in the treatment of eating disorders. Clinical Psychology & Psychotherapy, 28(3), 477–488. https://doi.org/10.1002/cpp.2622
Russell, H., Aouad, P., Le, A., Marks, P., Maloney, D., National Eating Disorder Research Consortium, Touyz, S., & Maguire, S. (2023). Psychotherapies for eating disorders: Findings from a rapid review. Journal of Eating Disorders, 11(1), 175. https://doi.org/10.1186/s40337-023-00886-w
Solmi, M., Wade, T. D., Byrne, S., Del Giovane, C., Fairburn, C. G., Ostinelli, E. G., De Crescenzo, F., Johnson, C., Schmidt, U., Treasure, J., Favaro, A., Zipfel, S., & Cipriani, A. (2021). Comparative efficacy and acceptability of psychological interventions for the treatment of adult outpatients with anorexia nervosa: A systematic review and network meta-analysis. The Lancet Psychiatry, 8(3), 215–224. https://doi.org/10.1016/S2215-0366(20)30566-6
Thompson, A., Calissano, C., Treasure, J., Ball, H., Montague, A., Ward, T., & Cardi, V. (2023). A case series to test the acceptability, feasibility and preliminary efficacy of AVATAR therapy in anorexia nervosa. Journal of Eating Disorders, 11(1), 181. https://doi.org/10.1186/s40337-023-00900-1