Emotionally Intelligent Chatbots in Mental Health: A Review of Psychological, Ethical, and Developmental Impacts

Use of emotionally intelligent chatbots is increasing in mental health settings to provide support by recognizing and reacting to users' emotions.This review has a closer look at 59 peerreviewed studies from 2017 to 2024, with a focus on systems like Woebot and Wysa.It maps out how affective computing, psychological frameworks like cognitive behavioral therapy (CBT), and human-computer interaction theories shape these systems.While there is early evidence of benefits like reduced anxiety and better emotional self-awareness, many issues remain unresolved.These include weak long-term evidence, cultural bias in emotion recognition, and potential overdependence on AI.We also highlight the risks of collecting and using emotional data without sufficient oversight.Based on this, we suggest future research should move toward multicultural, longer-term, and ethically grounded studies.The goal should be to create emotionally intelligent systems that support, not replace, genuine human connection, especially in vulnerable populations.