Privacy–Preserving Online Content Moderation: A Federated Learning Use Case

Users are exposed to a large volume of harmful content that appears daily on various social network platforms. One solution to users’ protection is developing online moderation tools using Machine Learning (ML) techniques for automatic detection or content fltering. On the other hand, the processing of user data requires compliance with privacy policies. In this paper, we propose a framework for developing content moderation tools in a privacy-preserving manner where sensitive information stays on the users’ device. For this purpose, we apply Diferentially Private Federated Learning (DP–FL), where the training of ML models is performed locally on the users’ devices, and only the model updates are shared with a central entity.