Privacy auditing framework reveals risks in federated learning for industrial security
Toward Realistic Privacy Risk in FL-Enabled OT Intrusion Detection Systems
Cryptography and Security
Summary
Industrial systems often use federated learning to train intrusion detection without sharing raw data, but this can still leak sensitive information. The authors introduce a new way to see how data flows during this learning process that older models miss. They propose a tool to check privacy risks by testing if attackers can guess if certain data was used to train the model. This approach is being tried out in a simulated smart grid to understand the risks better when such systems are used in real factories.
What this means in practice
- •For industrial cybersecurity teams: Implement privacy auditing tools in federated learning systems to detect and manage data leakage risks in operational technology environments.
- •For smart grid operators: Evaluate privacy vulnerabilities in intrusion detection models trained via federated learning deployed in smart grid infrastructures.
Tested on one dataset.
Authors
Assunta Imperatrice, Luigi ROmano
Abstract
In Operational Technology (OT) environments, Federated Learning (FL) is increasingly adopted to train Intrusion Detection Systems (IDSs) without sharing raw operational data across facilities. However, model updates may still leak sensitive information about industrial processes, creating privacy risks that remain largely invisible within current OT security architectures. Existing reference models, including the Purdue hierarchy, do not explicitly represent the FL learning lifecycle, making privacy leakage difficult to observe and evaluate. To address this gap, we introduce the ML Plane, a novel architectural abstraction that integrates FL workflows within OT systems, enabling visibility of learning-related information flows that are not explicitly represented in traditional OT architectures. Building on the proposed ML Plane, we outline a privacy auditing framework for FL-enabled OT systems. The framework uses Federated Membership Inference Attacks (FedMIA) as an initial assessment mechanism and is being instantiated within a Smart Grid testbed to support the future evaluation of privacy risks under realistic OT deployment constraints.