AI-Driven Personalized Positive Psychology Interventions for Enhancing User Psychological Resilience

Abstract

Digital positive psychology interventions (PPIs) are increasingly employed to enhance well-being and resilience, yet most existing systems remain static and insufficiently personalized. These limitations hinder adaptability, sustained engagement, and overall effectiveness. To address this gap, we propose an AI-driven personalized intervention framework that integrates user profiling, natural language processing, reinforcement learning-based feedback optimization, and explainable AI. The system dynamically tailors activities such as gratitude journaling, mindfulness practice, and strengths identification to each user's psychological profile. Empirical validation using synthetic datasets and a 10-week pilot study involving 210 participants demonstrates that the proposed framework outperforms both generic mobile applications and chatbot-based interventions. It achieved a 17.3% increase in life satisfaction (SWLS) and a 22.8% improvement in resilience (CD-RISC), with statistically significant results (p < 0.01). Ablation studies confirm the critical contribution of user profiling and adaptive feedback, while explainability enhances user trust and perceived autonomy. These findings suggest that integrating artificial intelligence with positive psychology offers a scalable, interpretable, and empirically effective pathway for promoting happiness and resilience-holding strong potential for deployment in educational, occupational, and clinical settings.