Combating Free-Riding in AIGC Service System: A Decentralized Reputation-Based Model Management Approach

Blockchain strengthens copyright protection for AI generated content (AIGC) by establishing a transparent and traceable management framework, which encourages model providers to participate in forming an AIGC service system, collectively driving the development of AIGC and offering high-quality, reliable content generation services to a broader audience. Despite blockchain enabling full traceability in AIGC generation, some malicious model providers may exploit free riding actions by intercepting user requests and forwarding them to other providers to save computational resources or pursue greater profits, and then returning the generated product to users directly or with slight modifications. Such behavior hinders users from accessing high-quality AIGC service resources and threatens the legitimate rights of honest model providers, which may ultimately diminish their enthusiasm to participate in the AIGC service system, disrupting the balance of value co-creation within the system. To combat such free-riding behaviors and ensure equitable benefit distribution among participants, we propose a decentralized reputation-based model management approach within the blockchain-enabled AIGC service system, reducing the probability of malicious service providers par ticipating while providing users with a reliable reference for model selection. Moreover, it protects the content generation through a timestamp-based watermark to prevent malicious alteration and unauthorized use, safeguarding the interests of all participants during the generation process and enhancing the reliability and security of the AIGC service system. Experimental results demonstrate that the proposed approach can effectively constrain and supervise model behaviors, successfully combating free-riding actions in the AIGC service system, and providing a reliable and intuitive reference for users in model selection.