ScaFedCrowd: A secure cross-system anonymous authentication and privacy-preserving task recommendation scheme for federated crowdsourcing

As a prominent area of research in the AI 2.0 era, crowdsourcing has garnered significant attention. Traditional crowdsourcing systems suffer from several drawbacks, including single points of failure, privacy breaches, and isolated resources. The development of decentralized, privacy-preserving federated crowdsourcing systems has consequently emerged as an inevitable trend. However, this shift also introduces new security challenges, such as ensuring participant privacy, enabling rational task recommendations, and establishing inter-system trust. To address these challenges, we propose ScaFedCrowd, a secure cross-system anonymous authentication and privacy-preserving task recommendation scheme for federated crowdsourcing. Our scheme utilizes blockchain with smart contracts as the underlying platform to manage the crowdsourcing process, facilitating secure inter-system collaboration. We propose an intra-system registration and authentication method that uses a trapdoor function and non-interactive zero-knowledge (NIZK) proofs to provide optional anonymity for users. Based on similarity measurement and ElGamal cryptography, our scheme achieves secure cross-system authentication and authorization between servers and workers. Furthermore, task recommendations based on simple hashing intersection technique provide a foundation for cross-system authentication, ensuring privacy while recommending the most suitable tasks for workers. Finally, a secure revocation and tracking mechanism ensures the protection of legitimate rights. Security analysis and simulation results demonstrate that ScaFedCrowd enhances effectiveness, security, and versatility.