This study aimed to identify and assess artificial intelligence (AI)-based mental health interventions based on multiple criteria, including effectiveness and accuracy, accessibility and usability, ethical and privacy concerns, psychological and human factors and integration with traditional healthcare.
Primary data were collected from mental health professionals, AI and healthcare researchers, patients/users, healthcare administrators and AI developers in Punjab, Pakistan. The Dynamic Grey Relational Analysis (DGRA) was used for ranking the interventions. Additionally, the Kruskal–Wallis test (KWT) was conducted to examine the significance of demographic variables such as gender, age, education, marital status and participant category. For comparative analyses the TOPSIS and the Analytical Ordinal Priority Approach (AOPA) models were used.
The results indicated that therapeutic effectiveness was the most significant factor. The KWT results showed no significant differences among demographic groups, suggesting that therapeutic effectiveness is consistently the most critical AI-based mental health intervention across different participant categories.
This study is the first of its kind to apply the multi-model framework to analyze AI-based mental health interventions. The findings provide valuable insights for policymakers, healthcare practitioners and AI developers to enhance the effectiveness and integration of AI-driven mental health solutions.
