Enhancing the Reliability of Academic Document Certification Systems with Blockchain and Large Language Models

Abstract

Verifying the authenticity of documents, whether digital or physical, is a complex and crucial challenge faced by a variety of entities, including governments, regulators, financial institutions, educational establishments, and healthcare services. Rapid advances in technology have facilitated the creation of falsified or fraudulent documents, calling into question the credibility and authenticity of academic records. Most existing blockchain-based verification methods and systems focus primarily on verifying the integrity of a document, paying less attention to examining the authenticity of the document’s actual content before it is validated and registered in the system, thus opening loopholes for clever forgeries or falsifications. This paper details the design and implementation of a proof-of-concept system that combines GPT-3.5’s natural language processing prowess with the Ethereum blockchain and the InterPlanetary File System (IPFS) for storing and verifying documents. It explains how a Large Language Model like GPT-3.5 extracts essential information from academic documents and encrypts it before storing it in the blockchain ensuring document integrity and authenticity. The system is tested for its efficiency in handling both digital and physical documents, demonstrating increased security and reliability in academic document verification.