---
type: "Concept"
title: "Federated unlearning"
description: "Removing a participant's influence from a trained model."
tags: [fl, governance, privacy, federated]
generated: { by: human:crpage, at: 2026-07-09T09:44:00Z }
status: stable
---

Methods to remove a client's contribution from an already-trained model — important for data-subject rights and exit. A maturing subfield and a sign that federated systems are governed systems, not just ML.