Collaborative generative artificial intelligence content identification and verification

Abstract

The Generative AI Content Verification Exchange systematically registers and stores content generated by AI and real people alike. Upon submission, the system categorizes content into distinct groups, then deconstructs it into multiple segments using various methods. Each segment is assigned a unique hash value, termed a 'part identifier,' ensuring individualized identification. This registration process, combining grouping, segmentation, and hashing, enhances content traceability and retrieval. The resulting database not only organizes generated content by groups but also allows for efficient and secure referencing of specific content segments. A content similarity score may be generated by comparing hash values against a large corpus of registered content. The similarity score is indicative of the likelihood, or not, of an input content being in part registered content.