Abstract: Systems and methods are provided for ensuring data security of distributed machine learning (ML) models (e.g., federated learning (FL)) through additive quantum encrypted noise. Large amounts of noisy data can be introduced, and the noisy data can be randomly generated by quantum simulations with, for example, three quantum gates. Systems and methods can deter dishonest participants that attempt to decode distributed private data through adversarial learning algorithms (e.g., generative adversarial networks (GANs)) by associating distributed data with encrypted quantum computations, which increase decoding difficulty due to highly complex quantum mechanics.