Evidence of a Cognitive Shift in AI Education: How Students Are Rethinking Human Intelligence?

Islem Rekik

Keywords: artificial intelligence; human intelligence; science of learning; AI education; cognition

TL;DR: This paper presents a longitudinal, exploratory analysis of how sustained exposure to generative AI reshapes students’ valuations of human and artificial intelligence.

Abstract: Perceptions of intelligence influence how learners evaluate and rely on artificial intelligence (AI) systems. Despite rapid advances in AI capabilities, little is known about how sustained exposure to AI tools affects students’ valuation of human (or natural) intelligence (HI) relative to artificial (or machine) intelligence. This \bluesky{} paper reports a longitudinal classroom response to a poll comparing the perceived importance of AI versus HI at the opening lecture of AI-focused courses between 2020 and 2026, spanning undergraduate and MSc programs in computer science. Responses from 471 students across technical (Machine Learning, Deep Graph Learning) and design-oriented (Design Thinking for AI) courses were analyzed. We identify four recurring phases: \textbf{(1) Hype, (2) Distrust, (3) Trust, and (4) Dependency}. Early measurements in 2020 slightly favored AI over HI. From 2024 onward, preferences consistently shifted toward human intelligence across all MSc cohorts, reaching 65\% (a 12 percentage-point increase from 2025) in a technical AI course and 90\% (a 36 percentage-point increase from 2025) in a design-oriented AI course by 2026. These observations suggest a gradual reappraisal of human intelligence as AI becomes a routine tool that may affect learners’ autonomy and epistemic agency. We conclude this perspective paper by offering introspective insights into a cognitive shift from favoring artificial intelligence toward prioritizing natural intelligence.

Paper Type: Blue Sky Paper

Submission Number: 59