Semihonest

In cryptography and secure multiparty computation, the semi-honest model (also called passive adversaries or honest-but-curious) assumes that participating parties follow the protocol but may try to learn additional information from the messages they receive. Under this model, a party’s internal state and transcripts may be analyzed to infer data beyond the intended input and output, even though the protocol’s behavior remains correct.

Formally, a protocol is secure in the semi-honest model if for every party there exists a probabilistic polynomial-time simulator that, given that party’s input and output, can generate a view indistinguishable from the party’s real view during execution. If such simulators exist for all parties, the protocol reveals no more information than the inputs and outputs. This simulator-based definition provides a rigorous baseline for privacy in secure computation.

Semi-honest security is weaker than malicious security, where participants may deviate from the protocol in arbitrary ways. Many protocols are designed first for semi-honest security and then transformed into actively secure or covertly secure versions using compiler techniques. The model is widely used as a tractable starting point and baseline in both two-party and multi-party computation.

Limitations include its failure to protect against purposeful deviation or collusion among participants. Consequently, semi-honest security is often insufficient in untrusted environments, though it remains a common building block and theoretical benchmark for secure computation.