🤖 AI Summary
To address identity traceability and privacy leakage risks arising from AI-driven eye-tracking in interactive education, this paper proposes a two-stage privacy-preserving framework. In Stage I, virtual identifiers, K-means clustering, and real-time anonymization are integrated to reduce identity re-identification accuracy to 63% while preserving diagnostic utility (classification accuracy: 99.3%). In Stage II, federated learning and fine-grained administrative access control are introduced, eliminating identity re-identification entirely and further improving classification accuracy to 99.40%. The framework strictly complies with GDPR and other regulatory standards. It represents the first approach to simultaneously guarantee high-accuracy neurodevelopmental disorder identification and strong identity unlinkability in educational settings. By enabling trustworthy governance of sensitive biobehavioral data—particularly eye-movement patterns—the framework provides a scalable, privacy-aware methodology for real-world edtech applications.
📝 Abstract
Eye-tracking technology can aid in understanding neurodevelopmental disorders and tracing a person's identity. However, this technology poses a significant risk to privacy, as it captures sensitive information about individuals and increases the likelihood that data can be traced back to them. This paper proposes a human-centered framework designed to prevent identity backtracking while preserving the pedagogical benefits of AI-powered eye tracking in interactive learning environments. We explore how real-time data anonymization, ethical design principles, and regulatory compliance (such as GDPR) can be integrated to build trust and transparency. We first demonstrate the potential for backtracking student IDs and diagnoses in various scenarios using serious game-based eye-tracking data. We then provide a two-stage privacy-preserving framework that prevents participants from being tracked while still enabling diagnostic classification. The first phase covers four scenarios: I) Predicting disorder diagnoses based on different game levels. II) Predicting student IDs based on different game levels. III) Predicting student IDs based on randomized data. IV) Utilizing K-Means for out-of-sample data. In the second phase, we present a two-stage framework that preserves privacy. We also employ Federated Learning (FL) across multiple clients, incorporating a secure identity management system with dummy IDs and administrator-only access controls. In the first phase, the proposed framework achieved 99.3% accuracy for scenario 1, 63% accuracy for scenario 2, and 99.7% accuracy for scenario 3, successfully identifying and assigning a new student ID in scenario 4. In phase 2, we effectively prevented backtracking and established a secure identity management system with dummy IDs and administrator-only access controls, achieving an overall accuracy of 99.40%.