Attack by Unlearning: Unlearning-Induced Adversarial Attacks on Graph Neural Networks

TL;DR

This paper reveals that privacy-preserving unlearning in graph neural networks can be exploited to perform stealthy adversarial attacks, causing significant accuracy degradation through carefully crafted node deletions.

Contribution

It introduces the concept of unlearning corruption attacks, demonstrating how adversaries can manipulate unlearning processes to compromise GNN performance.

Findings

Small, carefully designed unlearning requests can cause major accuracy drops.

Unlearning attacks are feasible across various GNN benchmarks and unlearning algorithms.

The attack exploits the legal and technical constraints of unlearning in GNNs.

Abstract

Graph neural networks (GNNs) are widely used for learning from graph-structured data in domains such as social networks, recommender systems, and financial platforms. To comply with privacy regulations like the GDPR, CCPA, and PIPEDA, approximate graph unlearning, which aims to remove the influence of specific data points from trained models without full retraining, has become an increasingly important component of trustworthy graph learning. However, approximate unlearning often incurs subtle performance degradation, which may incur negative and unintended side effects. In this work, we show that such degradations can be amplified into adversarial attacks. We introduce the notion of \textbf{unlearning corruption attacks}, where an adversary injects carefully chosen nodes into the training graph and later requests their deletion. Because deletion requests are legally mandated and cannot be denied, this attack surface is both unavoidable and stealthy: the model performs normally during training, but accuracy collapses only after unlearning is applied. Technically, we formulate this attack as a bi-level optimization problem: to overcome the challenges of black-box unlearning and label scarcity, we approximate the unlearning process via gradient-based updates and employ a surrogate model to generate pseudo-labels for the optimization. Extensive experiments across benchmarks and unlearning algorithms demonstrate that small, carefully designed unlearning requests can induce significant accuracy degradation, raising urgent concerns about the robustness of GNN unlearning under real-world regulatory demands. The source code will be released upon paper acceptance.