Advances Brain Research and Neuroscience
ISSN: 3071-2750 | DOI: 10.64030/3071-2750
Sora Pazer
The adoption of generative AI systems in particular large language models such as ChatGPT has expanded at a pace that extends far beyond academic or professional use cases. University students are increasingly turning to AI in emotionally challenging situations, using it as a conversational partner, reflective medium, and source of relief. The present study investigates whether this behavior can be conceptualized as a functional coping strategy within the framework of transactional stress theory [1]. A sample of N = 150 university students (M = 23.4 years, SD = 3.1) completed scales measuring AI coping, emotional relief, cognitive clarity, perceived stress, and ambivalence toward AI use. Results reveal significant positive associations between AI coping and emotional regulation (r = .62, p < .01) and cognitive clarity (r = .68, p < .01), alongside a negative association with perceived stress (r = -.31, p < .01). Multiple regression analysis identified AI coping as the strongest predictor of emotional relief (β = .58, p < .01, R2 = .42). Analysis of variance confirmed dose-dependent effects by usage intensity. At the same time, the data reflect structural ambivalence: AI is perceived as emotionally inauthentic and as an incomplete substitute for human relationships. Findings are discussed in relation to coping theory, digital health research, and the psychotherapy relationship literature.