Emotional support or problem-solving? A randomized controlled trial of function-specific AI Chatbots for digital stress in college students.

Digital stress has emerged as a salient risk factor for the mental health of emerging adults, yet targeted interventions remain limited. Generative AI chatbots offer a scalable approach to delivering context-sensitive supportive interactions that users may perceive as emotional or informational support. This randomized controlled trial tested two functionally specialized AI chatbots-emotional-support and problem-solving-in reducing digital stress, with mental health outcomes examined secondarily, among university students with elevated digital stress. Ninety-nine participants were randomized to emotional support, problem-solving, or waitlist control and completed a 14-day self-guided intervention. Results showed that both chatbots effectively increased their intended forms of perceived chatbot-delivered support and exhibited function-specific effects on different dimensions of digital stress. The emotional-support chatbot significantly reduced availability stress, approval anxiety, social comparison stress, and fear of missing out and improved secondary positive mental health. The problem-solving chatbot showed a distinct advantage in reducing online vigilance. Both interventions significantly alleviated anxiety symptoms, whereas effects on depressive symptoms were not significant. Findings suggest that functionally specialized AI chatbots may provide scalable support for digital stress, with effects depending on the alignment between support function and stress dimension.