AMCR: A Framework for Assessing and Mitigating Copyright Risks in Generative Models

2025-08-30

AI Summary

Text-to-image generative models often inadvertently reproduce copyrighted content, and existing prompt-filtering approaches fail to address latent copyright infringement. This paper proposes the first generation-process-oriented framework for proactive copyright-risk mitigation: (1) high-risk prompts are semantically reconstructed to avoid potential infringement; (2) an attention-guided cross-modal similarity analysis is introduced to precisely detect implicit copyright associations during image synthesis; and (3) an adaptive suppression module dynamically modulates copyright-sensitive features throughout the diffusion process. The framework preserves visual fidelity while substantially reducing infringement incidence—achieving an average 62.3% reduction in copyright leakage across multiple benchmarks compared to state-of-the-art methods. To our knowledge, this is the first approach enabling fine-grained, end-to-end controllability of copyright risk throughout the entire generative pipeline.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Natural Language Processing: Generation
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications

Abstract

Generative models have achieved impressive results in text to image tasks, significantly advancing visual content creation. However, this progress comes at a cost, as such models rely heavily on large-scale training data and may unintentionally replicate copyrighted elements, creating serious legal and ethical challenges for real-world deployment. To address these concerns, researchers have proposed various strategies to mitigate copyright risks, most of which are prompt based methods that filter or rewrite user inputs to prevent explicit infringement. While effective in handling obvious cases, these approaches often fall short in more subtle situations, where seemingly benign prompts can still lead to infringing outputs. To address these limitations, this paper introduces Assessing and Mitigating Copyright Risks (AMCR), a comprehensive framework which i) builds upon prompt-based strategies by systematically restructuring risky prompts into safe and non-sensitive forms, ii) detects partial infringements through attention-based similarity analysis, and iii) adaptively mitigates risks during generation to reduce copyright violations without compromising image quality. Extensive experiments validate the effectiveness of AMCR in revealing and mitigating latent copyright risks, offering practical insights and benchmarks for the safer deployment of generative models.

Problem

• Mitigating copyright risks in generative models
• Detecting partial infringements through similarity analysis
• Reducing copyright violations without compromising image quality

Innovation

• Systematically restructures risky prompts into safe forms
• Detects partial infringements via attention-based similarity analysis
• Adaptively mitigates risks during generation without quality loss