Harms and Ethical Risks of AI Detection Tools in Education: False Positives, Bias, Surveillance, and Student Rights

Since 2022, generative AI has become deeply embedded in educational contexts. This prompts universities and schools to deploy AI detection tools in the name of academic integrity. This paper accentuates that the dependence on AI detection tools reflects a broader tendency toward procedural outsourcing, whereby complex evaluative responsibilities are delegated to algorithms rather than addressed through pedagogical judgment and institutional governance. When deified as an instrument in academic integrity processes, AI detectors generate recurring harms: they misclassify legitimate student work, fail to reliably identify misconduct, and disproportionately disadvantage non-native English speakers and other linguistically marginalized learners. Their opacity compounds these issues, producing probabilistic scores without interpretable evidence and undermining procedural fairness. This paper also explains that AI detection tools flag non-native writing at higher rates, effectively penalizing linguistic variation and reinforcing dominant language norms. At the same time, while some bad-faith users can evade detection, high-achieving, original, and neurodivergent students are disproportionately accused. Routine reliance on these AI detection tools also normalizes surveillance, which shifts the educational relationship toward suspicion rather than trust. Framed within the United Nations' Sustainable Development Goals, particularly SDG 4 (Quality Education) and SDG 10 (Reduced Inequalities), this dependence on AI detection tools threatens both educational quality and equity. This paper contributes to scholarship on AI in education by demonstrating that detection-centered approaches to academic integrity are structurally flawed and to discussions on student rights by showing how these tools erode procedural fairness and disproportionately harm marginalized learners.