Application of AI Cognitive Deviations in Large Language Models (LLM): Implications of ChatGPT for Decision Aid and Sustainable Development in Higher Education

Abstract

This study aimed to identify the nature of cognitive failures in users of large language models such as ChatGPT, and to measure the level of these failures according to the gender variable (males and females) and the type of college (scientific and humanitarian colleges), while revealing possible statistical differences between them. To achieve this, the study relied on Broadbent's (1982) definition of cognitive failure. Cognitive failure scale for users of artificial intelligence applications to cover five interrelated areas, namely: distraction associated with the impact of the decision framework (Framing Bias), perception lapses associated with the impact of availability and stereotypical representation, memory errors associated with stereotypical representation, and functional motor failure associated with stereotypical representation and decision framework, in addition to integrating cases of cognitive failure with multiple cognitive biases. The results of the study showed that users of large language models are more prone to cognitive failure, as they suffer from difficulties in concentrating, absorbing accurate information, retrieving previous experiences, and implementing procedures accurately within a digital environment. The results also showed that there are no statistically significant differences between males and females, nor between students in scientific and human disciplines, indicating that exposure to cognitive failures is linked to the same use of artificial intelligence applications and not to demographic or specialized variables.

Ähnliche Arbeiten