Cognitive Debt in the ChatGPT Era: How Ethical and Emotional Use Shapes Cognitive Function in Emerging Adults

Abstract

Background:The accelerated use of large language models like ChatGPT has revolutionized human emotional and cognitive involvement, yet its neuropsychological implications remain poorly known. The present study proposes the concept of cognitive debt, the accumulated strain on attention, memory, and metacognitive control triggered by sustained AI engagement. This study investigated how distinctive patterns of ChatGPT involvement spanning usage frequency, emotional and cognitive engagement, and ethical reflection predict cognitive dysfunction across four user typologies: low–moderate, minimal/unhealthy, balanced–cognitive, and ethically reflective users. Method: This study employed a purposive sampling strategy within a web-based cross-sectional design to recruit 300 emerging adults (aged 18–25 years) from universities in Rawalpindi and Islamabad, Pakistan, between June 25 and July 12, 2025. Participants completed two standardized psychological instruments examining ChatGPT usage and cognitive dysfunction via an online survey administered on Google Forms. The survey link was disseminated through multiple digital platforms, including WhatsApp, Facebook, and official university email network to ensure broad accessibility and voluntary participation. Results: The results revealed that higher ChatGPT usage, specifically emotionally driven involvement, was associated with increased cognitive dysfunction, including impairments in memory, attention, and executive control across all user profiles, proposing that emotionally driven and impulsive interplay with generative AI diminishes executive control and heightens cognitive load. In contrast, ethical reflection indicated a mild protective effect against cognitive dysfunction. Moreover, females exhibited higher cognitive vulnerability than males, while males reported greater ChatGPT engagement and susceptibility to its cognitive effects as compared to females. Conclusions: The results explain two diverse cognitive stress pathways: (1) emotional compulsive engagement, described by affect-laden and impulsive AI use, and (2) reflective cognitive overload, where ethical contemplation paradoxically develops metacognitive load. These novel results improve the concept of cognitive debt, proposing that both over reflective and overreliant AI interactions could impair cognitive efficacy. The research highlights the urgency of establishing evidence-based digital ethical-use and literacy approaches to promote cognitively sustainable AI usage.

Ähnliche Arbeiten