🤖 AI Summary
To address copyright infringement risks posed by generative AI image models reproducing protected characters, this paper proposes a generalizable defense method grounded in quantitative originality assessment. Methodologically, it introduces (1) a novel, computable data originality estimation metric to objectively evaluate copyright risk in generated content; (2) PREGen—a unified framework integrating prompt rewriting and output generalization, effective under both explicit and implicit copyright-related prompts; and (3) originality-guided sampling to attenuate salient, copyright-protected character features during generation. Experiments demonstrate that PREGen reduces infringement rates by over 50% under explicit character references and suppresses infringing outputs to near-zero under implicit prompts. The approach significantly enhances compliance and generalization safety of AI-generated images without compromising visual quality or functional utility.
📝 Abstract
The rapid progress of generative AI technology has sparked significant copyright concerns, leading to numerous lawsuits filed against AI developers. Notably, generative AI’s capacity for generating images of copyrighted characters has been well documented in the literature, and while various techniques for mitigating copyright issues have been studied, significant risks remain. Here, we propose a genericization method that modifies the outputs of a generative model to make them more generic and less likely to imitate distinctive features of copyrighted materials. To achieve this, we introduce a metric for quantifying the level of originality of data, estimated by drawing samples from a generative model, and applied in the genericization process. As a practical implementation, we introduce PREGen (Prompt Rewriting-Enhanced Genericization), which combines our genericization method with an existing mitigation technique. Compared to the existing method, PREGen reduces the likelihood of generating copyrighted characters by more than half when the names of copyrighted characters are used as the prompt. Additionally, while generative models can produce copyrighted characters even when their names are not directly mentioned in the prompt, PREGen almost entirely prevents the generation of such characters in these cases. Ultimately, this study advances computational approaches for quantifying and strengthening copyright protection, thereby providing practical methodologies to promote responsible generative AI development.