Blockchain-Based copyright traceability scheme for generative artificial intelligence

Abstract

With the rapid development of Generative Artificial Intelligence (GAI), the explosive growth of Artificial Intelligence Generated Content (AIGC) brings great challenges to digital copyright protection. Featuring multimodality, large scale data and diverse creators, AIGC renders conventional centralized copyright solutions incapable of meeting storage efficiency, multiparty ownership management and full link traceability simultaneously. This paper proposes a blockchain-based multiparty collaborative evidence deposition and traceability scheme for AIGC. Integrated with IPFS for off-chain work storage, the scheme anchors only work identifiers on blockchain for evidence deposition. For rights confirmation, short signature technology enables ownership division via mult-party generation of holding right signatures embedded with contribution ratios. When copyright circulation takes place, chameleon hash supports trapdoor collision on work identifiers to implement ownership update. The scheme realizes full lifecycle AIGC management including rights confirmation and evidence deposition, copyright circulation and work traceability. Experimental results show that the scheme achieves satisfactory performance in security, functionality and efficiency, and provides feasible technical support for AIGC copyright protection.