Key points are not available for this paper at this time.
This study explores how individuals use ChatGPT, an artificial intelligence (AI) chatbot, for personal and emotional support. As generative AI tools become more widespread, users increasingly turn to them not only for information but also for comfort, reflection, and companionship. Using a survey-based approach, we examined five key areas: emotional use, perceptions of ChatGPT's emotional responsiveness, perceived risks of emotional attachment, factors influencing trust, and the impact on well-being. Findings indicate that many users engage with ChatGPT during emotionally vulnerable moments, appreciating its availability, nonjudgmental tone, and perceived empathy. While users recognize that ChatGPT does not experience real emotions, they often described its responses as thoughtful and supportive. Trust was shaped by response consistency, contextual sensitivity, and acknowledgment of limitations. For some, repeated positive interactions fostered a sense of comfort and self-reflection. However, participants expressed concerns about over-reliance, emotional misalignment, and blurred boundaries between human and AI. A minority reported feeling unsettled by generic or impersonal replies. Privacy and ethical considerations also emerged. These findings suggest a cultural shift in how people seek emotional support, highlighting both the potential and ethical complexity of emotionally responsive AI. Ensuring safe and supportive use will require responsible design, transparency, and ongoing user education.
Delello et al. (Fri,) studied this question.