Skip to content
#graph neural networks Conference Open access

Understanding Fairness–Privacy Interactions in Graph Neural Networks: A Literature Review, Taxonomy, and Open Challenges

Oct 2026 · Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society · Vol 9, pp. 2874-2884 · 0 citations · 41 references
Ethics and Social Impacts of AI

TL;DR

This paper proposes a two-dimensional orthogonal taxonomy that organizes existing work according to fairness objective and privacy target, and uses it to analyze dominant research patterns as well as the synergies and trade-offs between fairness and privacy.

Abstract

Graph Neural Networks (GNNs) are increasingly applied to socially consequential tasks, such as disease surveillance on contact graphs, and credit scoring and fraud detection on financial transaction networks. In these settings, biased or unfair predictions can restrict access to opportunities and resources, while privacy leakage can expose sensitive information, leading to substantial individual and societal harms. Consequently, fairness and privacy have become central concerns for trustworthy GNN deployment. Fairness-aware GNNs seek to reduce biased predictions across sensitive entities, whereas privacy-preserving GNNs aim to protect private graph information. However, these objectives are often interdependent: fairness interventions can affect privacy outcomes, while privacy mechanisms may impose additional constraints on fairness. This paper systematically reviews 19 studies to investigate the interplay between fairness and privacy in GNNs. We propose a two-dimensional orthogonal taxonomy that organizes existing work according to fairness objective and privacy target, and use it to analyze dominant research patterns as well as the synergies and trade-offs between fairness and privacy. This is the first review to provide a structured analysis of fairness–privacy interactions in GNNs and to identify open challenges and future research directions for trustworthy graph learning systems.

Read PDF

Similar papers

Book Open access Nov 2026

FairSEP: Harmonizing Multiple Fairness Metrics in Graph Neural Networks via Pareto Optimization

Group fairness in Graph Neural Networks (GNNs) aims to mitigate performance disparities across demographic groups. While existing methods primarily focus on Demographic Parity and Equal Opportunity, Predictive Equality (ΔFPR) remains under-explored, potentially leading to disparate error risks. Furthermore, these metho...

Kai-Xing Liang, Xue-Min Wang, Guo-Feng Xia et al. · 0 citations

From Network Inequality to Network Fairness: A Perspective on Responsible Decision-Making

It is shown that network biases are not inherently harmful or beneficial, and called for a holistic, networked approach to fairness that moves beyond static group categories and recognizes the dynamic, relational, and structural nature of inequality.

Lisette Espín-Noboa, Tina Eliassi-Rad, Pak-Hang Wong et al. · 0 citations
Open access Oct 2026

When Degree Anonymity Is Not Enough: Assessing Structural Re-identification Risk in Social Graphs

Removing identifiers and equalizing node degrees do not guarantee graph privacy, as network topology acts as a structural quasi-identifier. Evaluating rather than assuming privacy, this article frames structural re-identification risk assessment as a reproducible evaluation problem and presents a configuration-driven p...

C. J. P. de Brito, A. Vignatti, S. C. de Andrade · 0 citations
#federated learning Open access Oct 2026

Towards Fair and Robust Graph Federated Learning via Incentive Mechanisms

Graph Federated Learning (FL) allows multiple agents to collaboratively train a global graph model while preserving data privacy. However, current methods overlook the issue of low-quality or malicious data contributions, which can degrade the global model’s performance. This article addresses this gap by proposing a n...

Cheng-Lu Pan, Jia-Rong Xu, Carl Yang et al. · 0 citations

Related blog posts

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.