# Protocol for AI-based prediction of problematic digital technology use among Indian youth: a Centre for Advanced Research on Addictive Behaviours initiative

## TL;DR

This study aims to create an AI model to predict digital technology addiction in Indian youth, helping prevent mental health issues like stress and anxiety.

## Contribution

A novel AI-based predictive model for identifying Indian youth at risk of problematic digital technology use and associated psychological outcomes.

## Key findings

- An AI model will be developed and validated for early detection of digital technology use vulnerability.
- The model will incorporate demographic, psychological, and digital phenotype data for accurate predictions.
- Findings will support targeted interventions and digital-wellness policies in educational and community settings.

## Abstract

Artificial intelligence (AI) and machine learning have an important role in mental health research by helping to predict and prevent digital addiction and problematic digital technology use. These include behaviors linked with internet, smartphones, gaming, social media, gambling, over-the-top (OTT) platforms watching, pornography watching, shopping/ buying, and excessive screen time.

This multi-center study aims to develop and validate an AI-based predictive model to identify Indian youth at risk of problematic use of digital technology and associated psychological outcomes such as stress, anxiety, depression, and addiction. The study corresponds to one of the objectives of the Centre for Advanced Research on Addictive Behaviours (CAR-AB) initiative.

Students aged ≥12 years from schools and colleges across six Indian sites (New Delhi, Bhopal, Patna, Puducherry, Rishikesh, and Shillong) will be recruited. Data will be collected on demographic, psychological, behavioral, cognitive, socio-environmental, and digital phenotype correlates using validated instruments. Machine-learning models, including ensemble and deep-learning methods, will be trained, validated, and interpreted using explainable AI techniques.

The study will develop a validated and interpretable predictive model for early detection of vulnerability to problematic use of digital technology. Findings are expected to inform targeted interventions, guide digital-wellness policies, and support the integration of predictive tools within educational, work, and community settings.