A Survey of Defenses against AI-generated Visual Media: Detection, Disruption, and Authentication

2024-07-15

arXiv.org

AI Summary

To address risks—including disinformation, deception, and copyright infringement—arising from the misuse of deep generative models, this paper systematically surveys defense techniques against AI-generated visual content, focusing on detection, perturbation-based interference, and content authentication. We propose the first end-to-end, passive–active collaborative defense framework; introduce a cross-task transferable methodology taxonomy; and integrate trustworthiness evaluation dimensions—robustness, fairness, and verifiability. Through comprehensive literature analysis, formal modeling, and benchmark evaluation, we distill the technical evolution trajectory, identify core challenges (e.g., insufficient robustness and poor generalization), and outline future directions, including human–AI co-verification and cross-domain transferable defenses. (136 words)

Technology Category

• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Deep Generative Models & Autoencoders
• Humans and AI: Game Design — Procedural Content Generation & Storytelling

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Web Mining and Content Analysis: Web data quality in the era of algorithmically-generated content
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

Deep generative models have demonstrated impressive performance in various computer vision applications, including image synthesis, video generation, and medical analysis. Despite their significant advancements, these models may be used for malicious purposes, such as misinformation, deception, and copyright violation. In this paper, we provide a systematic and timely review of research efforts on defenses against AI-generated visual media, covering detection, disruption, and authentication. We review existing methods and summarize the mainstream defense-related tasks within a unified passive and proactive framework. Moreover, we survey the derivative tasks concerning the trustworthiness of defenses, such as their robustness and fairness. For each task, we formulate its general pipeline and propose a taxonomy based on methodological strategies that are uniformly applicable to the primary subtasks. Additionally, we summarize the commonly used evaluation datasets, criteria, and metrics. Finally, by analyzing the reviewed studies, we provide insights into current research challenges and suggest possible directions for future research.

Problem

• Surveying defense methods against malicious AI-generated visual media
• Reviewing detection, disruption and authentication for synthetic content
• Analyzing robustness and fairness of AI-generated media defenses

Innovation

• Systematic review of AI-generated media defenses
• Unified framework for passive and proactive protection
• Taxonomy based on methodological strategies for tasks