Verifiable AI-Assisted Learning Analytics on Blockchain Networks

Authors

  • A Renuka MAHGU, Dhaid Gaon, Block Pokhra , Uttarakhand, India raorenuka2@gmail.com Author

DOI:

https://doi.org/10.63345/

Keywords:

Learning Analytics, Blockchain, Federated Learning, Differential Privacy, Zero-Knowledge Proofs, Provenance, Verifiable Credentials, Consent Management, Governance, Auditability

Abstract

Learning analytics has matured into a strategic capability for institutions seeking to improve student success, personalize learning pathways, and optimize teaching practice. Yet the broader adoption of AI-assisted analytics is constrained by persistent trust and governance problems: opaque models, unverifiable data pipelines, fragmented consent management, and limited auditability of how insights are generated and used. This manuscript proposes a verifiable analytics architecture that anchors AI-assisted learning analytics on blockchain networks to provide endto-end integrity, provenance, and policy compliance. The approach integrates four pillars: (1) verifiable data provenance using standards such as xAPI/Caliper and content-addressed storage; (2) privacy-preserving model training and inference using federated learning and differential privacy; (3) cryptographic verification of analytics workflows using commitments, attestation, and zero-knowledge proofs; and (4) institutional governance through smart-contract-based consent, purpose limitation, Date of Publication: 12-06-2026 and verifiable credentials for outcomes. We describe the system model, threat assumptions, and a permissioned blockchain implementation that coordinates actors across learning management systems (LMS), devices, and analytics services. A methodology for evaluating performance, privacy risk, and verifiability is presented, along with illustrative results from a lab-scale prototype using synthetic data. The proposed design demonstrates how educational institutions can deliver actionable analytics while providing students and faculty with audit trails that independently verify the who, what, when, and how of analytics computation—without exposing raw personal data. We conclude with implications for policy, practice, and future research, including formal verification of smart-contract policies, standards alignment, and crossinstitution analytics marketplaces that preserve individual agency. 

Downloads

Download data is not yet available.

Additional Files

Published

12-06-2026

Issue

Section

Original Research Articles

How to Cite

Verifiable AI-Assisted Learning Analytics on Blockchain Networks . (2026). Scientific Journal of Artificial Intelligence and Blockchain Technologies, 3(2), Jun (50-62). https://doi.org/10.63345/

Similar Articles

21-30 of 129

You may also start an advanced similarity search for this article.