Personality, cognition, and AI use profiles of college students

This study seeks to better understand how different AI use profiles of college undergraduates vary in relationship to prompting and verification behaviors, personality variables, procrastination, and need for cognition. Using cluster analysis, three groups of students who approach AI in different ways were identified: power users, cautious users, and moderate users. Power users have high levels of trust, familiarity, and AI use for writing and low levels of distrust. Cautious users have low levels of trust, familiarity and AI use for writing, but high levels of distrust. Moderate users score at average levels across all of these variables. Power users tended to maximize their prompting for optimal results, but were less likely to verify AI output. They were more extraverted and agreeable. In contrast, cautious users were less likely to modify prompts for maximal results but were the most likely to verify. They were highest in openness and were higher in need for cognition than moderate users. The study demonstrates that subgroups of AI users can be identified and these groups differ in both cognitive style and personality which affects how they use AI in their academic lives.