SlideChain: Semantic Provenance for Lecture Understanding via Blockchain Registration
Md Motaleb Hossen Manik, Md Zabirul Islam, Ge Wang
TL;DR
SlideChain tackles the challenge of verifying AI-derived semantic content in educational slides by anchoring per-slide semantic provenance on an Ethereum-compatible blockchain. It introduces a four-model semantic extraction pipeline, normalizes and unifies concepts and relational triples, and stores compact cryptographic hashes on-chain to enable tamper-evident audits and reproducibility. The study reveals substantial cross-model semantic disagreement and reveals that single-model provenance misses large portions of the semantic content; it also demonstrates predictable gas usage, linear scalability, and perfect reproducibility across runs. The approach provides a practical, scalable foundation for auditability and long-term integrity in AI-assisted STEM education, enabling provenance-guided evaluation, drift detection, and human-in-the-loop oversight.
Abstract
Modern vision--language models (VLMs) are increasingly used to interpret and generate educational content, yet their semantic outputs remain challenging to verify, reproduce, and audit over time. Inconsistencies across model families, inference settings, and computing environments undermine the reliability of AI-generated instructional material, particularly in high-stakes and quantitative STEM domains. This work introduces SlideChain, a blockchain-backed provenance framework designed to provide verifiable integrity for multimodal semantic extraction at scale. Using the SlideChain Slides Dataset-a curated corpus of 1,117 medical imaging lecture slides from a university course-we extract concepts and relational triples from four state-of-the-art VLMs and construct structured provenance records for every slide. SlideChain anchors cryptographic hashes of these records on a local EVM (Ethereum Virtual Machine)-compatible blockchain, providing tamper-evident auditability and persistent semantic baselines. Through the first systematic analysis of semantic disagreement, cross-model similarity, and lecture-level variability in multimodal educational content, we reveal pronounced cross-model discrepancies, including low concept overlap and near-zero agreement in relational triples on many slides. We further evaluate gas usage, throughput, and scalability under simulated deployment conditions, and demonstrate perfect tamper detection along with deterministic reproducibility across independent extraction runs. Together, these results show that SlideChain provides a practical and scalable step toward trustworthy, verifiable multimodal educational pipelines, supporting long-term auditability, reproducibility, and integrity for AI-assisted instructional systems.