Malicious Adversary

A malicious adversary may deviate from the protocol, submit malformed inputs, manipulate outputs, or selectively abort.

Practical Example

A participant in federated learning sends poisoned updates while secure aggregation makes individual update inspection difficult.

Good Fit

Use this model when parties have competing incentives, financial stakes, or weak governance.

PET Implications

Prefer malicious-secure protocols, input validation, robust aggregation, audit logs, rate limits, and incident response plans.

Common Gap

Many prototypes claim privacy against curious parties while product teams assume protection against malicious ones.