After winning 4 awards for my poster at ABCT, my research paper has finally published in npj digital medicine Nature Portfolio! Association for Behavioral and Cognitive Therapies🎉
What is the efficacy of AI-delivered CBT in reducing anxiety and depressive symptoms?
With AI-powered therapy chatbots becoming increasingly popular, we explored this question by reviewing 16 RCTs (2,875 participants) testing AI-delivered CBT for anxiety and depression.
What we found:
→ Anxiety: only 1 of 14 interventions showed a significant benefit in intention-to-treat analyses
→ Depression: only 4 of 16 showed a significant benefit in intention-to-treat analyses
→ More advanced AI with more personalisation nor added human support led to better outcomes
→ Preliminary signals/ questions to consider: Third-wave approaches (e.g. ACT, mindfulness) and younger users showed early promise, while the few studies in older adults showed little benefit.
Limitations: Every study had “some concerns” or high risk of bias, mostly because of high dropout. Only one trial used generative AI in our search.
đź’ˇ My biggest takeaway?
Just because AI is becoming more sophisticated doesn't necessarily mean it's becoming more effective at supporting our mental health. AI has an enormous potential to make mental healthcare more accessible, but we need stronger scientific evidence (better study quality and lower attrition) and user-centred design before we can fully realize that potential.
Read the paper here: https://www.nature.com/articles/s41746-026-02744-w
Also a huge thank you to my brilliant co-authors David Stuckler Ya Yambao Yang Lucia Wang for making this happen.
#DigitalHealth #MentalHealth #ArtificialIntelligence #CBT #SystematicReview #npjDigitalMedicine
My first author paper at npj digital medicine
After winning 4 awards for my poster at Association for Behavioral and Cognitive Therapies conference, my research paper has finally published in npj digital medicine, Nature Portfolio!