This study introduces a four-pillar student-activity framework (Studying and Learning, Research and Projects, Personal and Career Development, and Campus and Community Life) to analyze AI’s ethical challenges in higher education. Drawing on peer-reviewed sources from 2022 to 2025, we identify recurring risks across pillars: academic integrity, privacy/data protection, bias/fairness/equity, student agency/(de)skilling, and governance gaps. We distill three cross-pillar principles: disclosure plus process evidence (e.g., prompt/version logs), privacy-by-design, and proportionality and equity/fairness scaffolds (institutional access, bias audits, and multilingual support). These translate into actionable strategies for assessment redesign, research supervision, career services, and campus operations. The framework unifies fragmented discourse, supports institutional decision making, and reveals gaps for longitudinal and causal research. It demonstrates that responsible AI use emerges when processes are visible, data practices are proportionate, and access is equitable, amplifying human learning without eroding trust or integrity.
Madleňák et al. (2026) studied this question.