Neuroethics of university assessment based on learning analytics and predictive models

Introduction: Learning analytics and predictive models have become consolidated tools that enable the early identification of academic risk, the personalization of feedback, and data-driven decision-making. However, the growing reliance on algorithmic assessments, automated predictions, and behavioral data raises ethical concerns related to cognitive privacy. These technologies challenge traditional principles of fairness. The objective of this study is to analyze the ethical challenges associated with the use of learning analytics and predictive models in university assessment.Method: A quantitative, non-experimental, cross-sectional study with a descriptive–correlational scope was conducted. The participants included 132 university students and faculty members from health and education programs. A structured questionnaire of 30 Likert-type items (α = 0.91) was applied to measure algorithmic assessment practices, acceptance of predictive models, and perceptions of neuroethical risks.Results: A moderate level of acceptance of assessment based on learning analytics was identified, while the perception of risks related to cognitive privacy and algorithmic fairness reached high levels. Significant correlations were found between technological familiarity and acceptance, as well as negative relationships between acceptance and the perception of ethical risk.Conclusions: Learning analytics and predictive models offer significant opportunities to improve university assessment; however, their ethical implementation requires transparency, explainability, governance of educational data, and safeguards to protect students’ mental autonomy.