Adversarial Machine Learning for Social Good: Reframing the Adversary as an Ally

Aug 4, 2026

Upload by Ezra

Deep neural networks (DNNs) underpin recent advances in machine learning, yet research shows they are vulnerable to adversarial examples that perturb inputs to induce incorrect model decisions. Embedded bias and limited explainability can further enable anti-social AI applications, while large language models such as ChatGPT and GPT-4 amplify the risk at scale. This review introduces AdvML for Social Good (AdvML4G) as a field that repurposes adversarial learning to develop pro-social applications, calling for coordinated efforts among regulators, practitioners, and researchers.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages22

Reading Time55 min

AccessFree to download

Introduction

AdvML vulnerabilities and adversarial examples

Embedded bias and explainability concerns

AdvML for Social Good (AdvML4G) overview

Taxonomy and comparative analysis (AdvML vs AdvML4G)

Social good concepts and aspects

Motivations and intersection with ML4G

Challenges and open research issues

©2024 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, forresale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works. IEEE Transactions on Artificial Intelligence Adversarial Machine Learning for Social Good: Reframing the Adversary as an Ally | Journal: | IEEE Transactions on Artificial Intelligence | | --- | --- | | Manuscript ID | TAI-2023-Oct-R-01022 | | Manuscript Type: | Review Article | | Date Submitted by the Author: | 03-Oct-2023 | | Complete List of Authors: | Al-Maliki, Shawqi ; Hamad Bin Khalifa University
Qayyum, Adnan; Information Technology University,
Ali, Hassan; Information Technology University, Department of Electrical Engineering
Abdallah, Mohamed; Hamad Bin Khalifa University
Qadir, Junaid; Qatar University
Huang, Dinh ; University of Technology Sydney, FEIT
Niyato, Dusit; Nanyang Technological University, School of Computer Science and Engineering
Al-Fuqaha, Ala; Hamad Bin Khalifa University | | Keywords: | Adversarial learning, Accountable artificial intelligence, Artificial intelligence in computer security | | | | Page 1 of 20 IEEE Transactions on Artificial Intelligence JOURNAL OF IEEE TRANSACTIONS ON ARTIFICIAL INTELLIGENCE, VOL. 00, NO. 0, OCTOBER 2023 1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 Adversarial Machine Learning for Social Good: Reframing the Adversary as an Ally Shawqi Al-Maliki, Adnan Qayyum, Hassan Ali, Mohamed Abdallah, Senior Member, IEEE, Junaid Qadir, Senior Member, IEEE, Dinh Thai Hoang, Senior Member, IEEE, Dusit Niyato, Fellow, IEEE, Ala Al-Fuqaha∗ , Senior Member, IEEE Abstract—Deep Neural Networks (DNNs) have been the driving force behind many of the recent advances in machine learning. However, research has shown that DNNs are vulnerable to adversarial examples—input samples that have been perturbed to force DNN-based models to make errors. As a result, Adversarial Machine Learning (AdvML) has gained a lot of attention, and researchers have investigated these vulnerabilities in various settings and modalities. In addition, DNNs have also been found to incorporate embedded bias and often produce unexplainable predictions, which can result in anti-social AI applications. The emergence of new AI technologies that leverage Large Language Models (LLMs), such as ChatGPT and GPT-4, increases the risk of producing anti-social applications at scale. AdvML for Social Good (AdvML4G) is an emerging field that repurposes the AdvML bug to invent pro-social applications. Regulators, practitioners, and researchers should collaborate to encourage the development of pro-social applications and hinder the development of anti-social ones. In this work, we provide the first comprehensive review of the emerging field of AdvML4G. This paper encompasses a taxonomy that highlights the emergence of AdvML4G, a discussion of the differences and similarities between AdvML4G and AdvML, a taxonomy covering social good-related concepts and aspects, an exploration of the motivations behind the emergence of AdvML4G at the intersection of ML4G and AdvML, and an extensive summary of the works that utilize AdvML4G as an auxiliary tool for innovating pro-social applications. Finally, we elaborate upon various challenges and open research issues that require significant attention from the research community. Impact Statement—Adversarial Machine Learning (AdvML) is a research field that harnesses adversarial attacks to demonstrate and exploit models’ vulnerabilities. While these v

FAQs

What are adversarial examples and why do they matter in DNN-based systems?

Adversarial examples are input samples whose features are perturbed to force DNN-based models to make errors. They highlight vulnerabilities that can be exploited in harmful or unintended ways.

How does AdvML4G use adversarial machine learning for social good?

AdvML4G repurposes adversarial learning to create pro-social applications. It reframes the adversary as an ally to mitigate anti-social or exploitative uses of machine learning.

Why is AdvML4G increasingly relevant with the rise of large language models?

New AI technologies leveraging large language models increase the risk of producing anti-social applications at scale. This motivates coordinated development of pro-social systems and research into misuse prevention.

Aug 4, 2026

Upload by Ezra