Zero-Knowledge Consensus Protocols for Privacy-Preserving Multi-Agent Systems

In applications like tax fraud detection (Alluri, Source 3, Heuristic 1) and collaborative robotics, agents often need to coordinate without exposing sensitive information. Most current MAS protocols assume transparent state or intent sharing, which is increasingly untenable. This research proposes a family of consensus and coordination protocols leveraging zero-knowledge proofs (ZKPs) and secure multiparty computation, enabling agents to “prove” compliance with consensus rules or task constraints without disclosing their local states. Unlike traditional cryptographic applications, here the focus is on real-time, scalable, and lightweight ZKP constructions tailored for resource-constrained MASs. This would be a substantial leap beyond “secure consensus under attacks” frameworks (Han et al., Source 4, Heuristic 2), addressing both privacy and correctness. The impact is enabling MASs for sensitive domains (healthcare, finance, defense) to collaborate securely, preserving both privacy and trust.

References:

1. Multi-Agent Systems for Coordinated Fraud Detection in Tax Networks. Ravi Kiran Alluri (2025). International Journal of Multidisciplinary Research and Growth Evaluation.
2. Secure Sampled-Data Consensus of Multi-Agent Systems Under Asynchronous Deception Attacks With Application to Unmanned Surface Vehicles. Shengzhi Han, Hong Zhu, Qishui Zhong, K. Shi, O. Kwon (2024). IEEE transactions on intelligent transportation systems (Print).