Challenging the Honest-but-Curious Assumption: MPC Under Strategic, Partially Trusted Coalitions

Most MPC work (e.g., Benhamouda et al., 2018; Woods et al., 2022) assumes either honest-but-curious, fully malicious, or statically defined adversary sets. However, in real-world collaborations (such as supply chains or data marketplaces), trust relationships are fluid—parties might cooperate temporarily, or act strategically without fully colluding. This research would formalize and construct MPC protocols that recognize and adapt to such “gray zone” adversary models, possibly drawing from game theory, reputation systems, and dynamic trust metrics. The protocols could, for example, adapt their verification mechanisms or secret-sharing schemes based on observed or predicted coalition behaviors, and log “trust events” for auditability. This direction challenges the static assumptions of most current MPC, aiming for protocols that better map onto messier, more realistic trust environments—potentially unlocking new applications in finance, healthcare, and cross-border data sharing.

References:

1. Secure Comparisons of Single Nucleotide Polymorphisms Using Secure Multiparty Computation: Method Development. Andrew Woods, Skyler T. Kramer, Dong Xu, Wei Jiang (2022). JMIR Bioinformatics and Biotechnology.
2. Two-Round Adaptively Secure Multiparty Computation from Standard Assumptions. Fabrice Benhamouda, Huijia Lin, Antigoni Polychroniadou, Muthuramakrishnan Venkitasubramaniam (2018). Theory of Cryptography Conference.