“

Dartmouth College researchers recently conducted a groundbreaking clinical trial utilizing a therapy chatbot powered by generative AI. The findings, published in NEJM AI, a journal from the publishers of the New England Journal of Medicine, revealed significant improvements in participants’ symptoms with the use of the innovative software.

Key Findings:

• The trial involved 106 individuals diagnosed with major depressive disorder, generalized anxiety disorder, or an eating disorder.
• Participants engaged with the chatbot, known as Therabot, through a smartphone app by responding to prompts or initiating conversations.
• Those with depression experienced a 51% reduction in symptoms, leading to notable improvements in mood and overall well-being.
• Generalized anxiety participants reported a 31% average reduction in symptoms, with many shifting to milder anxiety levels.
• Users at risk of eating disorders showed a 19% average reduction in body image and weight concerns, outperforming the control group.

The results underscore the potential of AI-powered therapy to offer real-time support to individuals lacking immediate access to mental health professionals. While acknowledging the necessity of clinician oversight, the researchers believe that AI could bridge gaps in mental health services.

Lead author Nicholas Jacobson, an associate professor at Dartmouth’s Geisel School of Medicine, emphasized the AI-assisted approach’s potential for meaningful clinical benefits. Given the shortage of mental health providers in the U.S., Jacobson envisions AI offering support to a broader population in need of mental health services.

Therabot Development:

• Developed at Dartmouth since 2019, Therabot underwent continual consultation with mental health professionals.
• The app generates natural, open-ended responses rooted in evidence-based psychotherapy practices.
• Therabot demonstrated its safety protocols by prompting users to seek help in emergencies like suicidal ideation.

The clinical trial revealed that Therabot users experienced significant symptom reductions, surpassing statistical significance thresholds in just eight weeks. Users engaged with Therabot for an average of six hours, reflecting the software’s potential impact on user engagement and adherence.

The study also highlighted the emergence of a therapeutic alliance between users and the chatbot, mirroring the rapport often seen in traditional therapy settings. Users expressed comfort and trust in interacting with the software, leading to open and honest conversations, even during vulnerable times.

Challenges and Future Prospects:

• Michael Heinz, the study’s first author, stressed the importance of further research into the safety of using generative AI in mental health contexts.
• He added that diligent oversight by mental health experts is crucial in ensuring the software’s safety and efficacy.
• Despite the promising results, the researchers are cautious about the use of AI in high-risk mental health scenarios.

As the field of AI-assisted therapy advances, ongoing supervision and robust safety measures will be essential to safeguard users’ well-being. The Therabot trial exemplifies the potential of generative AI to enhance mental health care delivery and accessibility while emphasizing the importance of safety and efficacy in AI applications within this critical domain.