BCAIGC: Trusted, Efficient, and Reasonable AIGC Service and Trading Framework in Web3

Abstract

Artificial Intelligence Generated Content (AIGC) services have shown tremendous potential in digital content creation. However, AIGC service, trading, and product ownership have not been adequately protected due to the lack of a reliable third-party verification platform. Although blockchain-based AIGC management solutions have been proposed, they still suffer from several overlooked issues, such as performance bottlenecks in blockchain systems and repeated sales of AIGC products. Motivated by these limitations, we propose a trusted, efficient, and reasonable AIGC service and trading framework called BC-AIGC by combining identifiers, state channels, and game theory with Web3. Specifically, we first propose an AIGC management scheme based on identifiers and blockchain that enables AIGC ownership to be traced. Then, an AIGC service and trading method based on state channels is proposed, allowing most processes to be completed off-chain and resolving non-cooperation through dispute handling mechanisms, thereby improving efficiency while ensuring security. Furthermore, to find the optimal strategy, we model the interactions among AIGC service participants as a Stackelberg game and prove the existence of a unique Stackelberg equilibrium. Finally, the system prototype and numerical simulations demonstrate the superiority of the proposed BC-AIGC.