Using Conversational AI to Teach Growth Mindset Skills to Youths in India

Overview

This study aims to test whether a brief digital intervention using conversational AI can improve mental health outcomes among school-aged youth (grades 6-8) in India, where most young people with mental health issues do not receive treatment. The intervention teaches "growth mindset", the belief that skills and abilities can improve with effort, via a 45-minute interactive conversation with an AI chatbot (spread across two class periods). A randomized controlled trial with approximately 430 students at a private, English-medium school in Bangalore will evaluate whether interacting with the chatbot can reduce symptoms of anxiety and depression while enhancing growth mindset beliefs. Half the students will receive the full intervention at baseline, while the other half (control group) will complete a usual school assignment; after 7 weeks, the control group will receive a shortened (15-minute) version of the growth mindset intervention. Participants will complete 10-minute surveys at baseline, 3 weeks, and 7 weeks, after which all students will receive a printed booklet with all the information from the chatbot. This study represents one of the first randomized controlled trials evaluating conversational AI as a brief digital intervention for youth mental health.

Globally, 300 million youths have a diagnosable mental disorder and more reside in India than any other country. Yet, in India, where this study is based, most youths affected by mental health issues do not receive treatment. Symptoms are especially prevalent during high school, when many students show signs of depression and anxiety. Given myriad challenges in accessing mental healthcare, some have turned to conversational AIs such as ChatGPT and Claude for personalized support. However, there are few rigorous tests of whether such tools can improve mental health. This study aims to bridge this gap by combining conversational AI with evidence-based intervention for youth mental health-specifically, a growth mindset intervention, which promotes the belief that abilities are not innate but can be improved with effort. It is well known that growth mindset interventions can improve academic performance. In addition, researchers have shown that brief, computer-guided growth mindset training can reduce symptoms of anxiety and depression that sustain 3-9 months post-intervention, and not only in the United States. However, until now, these interventions have been delivered via static, standardized websites that few people use. The intervention considered is a 45-minute interactive lesson on growth mindset, embedded within a conversational AI. Similar to prior interventions, the content includes modules on understanding growth mindset, its basis in neuroplasticity, and roleplaying scenarios. Each module includes interactive discussions with the AI (e.g., "in your words, what is a growth mindset?", "in the above scenario, how could a growth mindset apply?") The specific aim of the study is to test whether a brief, AI-enabled growth mindset intervention offers measurable benefits for school-aged youth in India. A randomized controlled trial with about 430 students (grades 6-8) will measure whether the AI-enabled training leads to changes in growth mindset beliefs and skills as well as symptoms of anxiety and depression, relative to a control group that receives a shortened intervention after 7 weeks. Brief, 10-minute questionnaires will be administered to all participants at baseline, 3 weeks, and 7 weeks. At the conclusion of the trial, all students will receive a printed booklet with content from the full intervention. The study will take place in an English-medium private school in Bangalore, India, which serves a diversity of students. The growth mindset intervention may also benefit academic performance, and the school partner, at its discretion, may be able to integrate it into part of their standard curriculum following the completion of the study. However, the specific research question is whether the chatbot interaction yields measurable improvements in youth mental health. Given the preponderance of AI chatbots aiming to address various aspects of mental health, such measurement is as important as it is rare. This study will contribute to the growing evidence base on how technology can inform and support youth mental health.