Mitigating Information Leakage Risks in Secure Multiparty Computation through Function Hiding

Secure multiparty computation (SMPC) allows a joint computation on private data, but the majority of the existing protocols assume implicitly that the computation being performed is public and nonsensitive. In practice, though, computation logic frequently incorporates proprietary strategy or sensitive rule of decision, and its exposures constitute a significant though neglected leak of information. The prevailing SMPC models can mainly provide input confidentiality and accuracy, but the protocol level leakage due to observable protocol behavior is not well tackled. This paper attempts to fill this gap by introducing a (Function-Hiding Secure Multiparty Computation) FH-SMPC framework modifying the SMPC workflow to incorporate encapsulation of functions and regular patterns of execution directly. The suggested design hides structural and semantic attributes of the considered function and maintains the correctness, scalability, and compatibility with the conventional SMPC primitives. An explicit security analysis provides indistinguishability of functions along with classical input privacy. Experimental evaluation shows that FH-SMPC reduces transcript-based function distinguishability by over 85%, achieving an average divergence of 0.028 compared to 0.214 for baseline SMPC, with an execution latency increase limited to approximately 12% and no asymptotic growth in communication overhead. By treating computation logic as a confidential asset, FH-SMPC advances secure collaborative computation and provides a practical foundation for privacy-sensitive applications in cloud analytics and distributed decision systems.