Global Study: Openness and Conscientiousness Predict Gen-AI Use in Higher Education

Who uses Gen-AI most in higher education? A 4-country study breaks it down

A cross-national team from Türkiye, India, South Korea, and Kuwait surveyed 1,016 students across engineering, medicine, and other fields to profile how personality traits influence Gen-AI use in learning. All participants had prior experience with tools like ChatGPT, Bing AI, or Jasper. The findings were published in Scientific Reports.

The strongest predictors

Two traits stood out as the best predictors of frequent Gen-AI use: openness to experience (originality, curiosity, artistic sensitivity, inquisitiveness) and conscientiousness (hardworking, self-regulated, careful, achievement-oriented). Students high in either were the most active users.

Extraversion also nudged usage upward, but less than the two core traits. Neuroticism (anxiety, emotional instability, self-consciousness, frustration) lowered the odds of use. Agreeableness showed no meaningful link. For quick context on the traits, see the APA overview of the Big Five factors here.

Gender patterns worth noting

Among women, conscientiousness was the stronger driver of Gen-AI use. Among men, openness had the larger effect. Extraversion encouraged use in both groups but had a stronger influence for women.

Neuroticism discouraged use more for women than men. Agreeableness again showed no practical connection in either group.

What this means for educators, learning designers, and edtech teams

• Lean into curiosity: Offer flexible, exploratory prompts and sandboxed tools for students high in openness.
• Support goal-driven learners: Provide clear rubrics, checklists, and accountability structures to align with conscientious users.
• Use interaction to drive engagement: Live demos, peer critique, and collaborative prompts can help extraverted students participate more.
• Reduce friction for anxious learners: Provide low-stakes practice, transparent grading rules for AI use, and quick-start guides to lower uncertainty.
• Avoid assumptions based on agreeableness: It's not a reliable signal for AI adoption.
• Measure ethically: If you collect personality data, keep it opt-in, minimal, and separate from grading. Be explicit about purpose and retention.

Implementation checklist for the next term

• Define allowed AI use per assignment (use cases, boundaries, citation rules) and brief students on why and how.
• Offer two paths for core tasks: a structured track (stepwise prompts, templates) and an exploratory track (open-ended briefs, tool choice).
• Provide micro-guides: 5-10 minute walkthroughs for setup, prompt examples, and common pitfalls.
• Create an "AI practice zone": low-stakes exercises that let students test tools without grade risk.
• Train instructors and TAs on consistent policy enforcement and quick diagnostics for AI misuse.
• Monitor signals that matter: task quality, time-on-task, resubmission rates, and integrity incidents-then adjust prompts and supports.

Scope and limits

The sample was 17-28 years old and largely undergraduate. It spanned four countries and multiple disciplines, which increases relevance but doesn't replace local testing. Treat these trait effects as directional signals to inform design, not fixed rules.

Building program-wide capability? Explore curated AI course lists by role at Complete AI Training.