AI Chatbot JEM Boosts Body Image and Mood in Real-World Study

In an era where digital devices are frequently blamed for eroding self-esteem, an automated conversational agent is demonstrating that technology can also provide an accessible lifeline. A study published in JMIR Human Factors reveals that JEM, a rule-based chatbot designed to address body dissatisfaction and disordered eating, delivered immediate, real-world improvements in user mood and body image across Australia and Canada.

Immediate Relief Through Digital Microinterventions

Participants who engaged with JEM reported measurable, within-session boosts in confidence, emotional well-being, and body appreciation after completing brief interactive exercises. Rather than asking users to commit to lengthy therapeutic modules, the chatbot provides bite-sized exercises that users can complete on demand when facing acute moments of distress.

Public reception was overwhelmingly positive. User satisfaction consistently exceeded 76% across both trial regions, averaging 76.1 in Australia and 78.8 in Canada. Furthermore, the platform’s ease of use received an “excellent” rating on the System Usability Scale, scoring an average of 86.5 in Australia and 89.5 in Canada, according to the researchers.

Addressing a Critical Gap in Eating Disorder Care

Concerns surrounding body image, disordered eating habits, and clinical eating disorders represent urgent public health challenges worldwide. Despite the severe psychological burden these conditions impose, the vast majority of individuals experiencing distress never receive evidence-based clinical care due to systemic barriers such as prohibitive costs, long waiting lists, and pervasive social stigma.

To help bridge this treatment gap, researchers co-designed JEM with community stakeholders. The platform was built to deliver accessible psychoeducation and evidence-based therapeutic exercises directly into the hands of people who might otherwise navigate their struggles in isolation.

Testing the Tool in Everyday Settings

Moving beyond preliminary laboratory testing, the research team tracked real-world engagement over a six-month deployment. A total of 230 users aged 13 to 64 completed an overall evaluation survey, including 122 respondents in Australia and 108 in Canada.

To capture immediate shifts in user experience, participants were invited to complete brief surveys before and after specific microinterventions, with participation counts ranging between 75 and 276 depending on the exercise. The assessments used validated visual analog scales to measure real-time changes in anxiety, depression, happiness, and confidence, as well as perceptions of physical attractiveness, body shape, and body size satisfaction.

Who JEM Reaches and What Comes Next

Demographic analyses indicated that the chatbot primarily attracted young adult cisgender women and nonbinary individuals in both countries—populations that face elevated risks of body dissatisfaction and eating pathology. Within-session analyses revealed that the exercises generated benefits ranging from very small to large effect sizes across various mood and body image metrics.

While these initial real-world findings demonstrate that automated tools can offer rapid emotional relief, the study’s authors emphasize that further investigation is warranted. Future studies will need to explore how repeated engagement with chatbots like JEM translates into long-term mental health benefits and whether digital microinterventions can effectively complement formal eating disorder prevention and care.

---