Summary of Rethinking Artistic Copyright Infringements in the Era Of Text-to-image Generative Models, by Mazda Moayeri et al.
Rethinking Artistic Copyright Infringements in the Era of Text-to-Image Generative Models
by Mazda Moayeri, Samyadeep Basu, Sriram Balasubramanian, Priyatham Kattakinda, Atoosa Chengini, Robert Brauneis, Soheil Feizi
First submitted to arxiv on: 11 Apr 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary Recent advancements in text-to-image generative models, such as Stable Diffusion, have raised concerns about copyright infringement in the art world. Understanding how these models copy artistic style is more complex than just duplicating a single image. In this paper, we reformulate the problem of “artistic copyright infringement” to a classification problem over image sets, instead of probing image-wise similarities. We introduce ArtSavant, a practical tool to determine an artist’s unique style by comparing it to a reference dataset of works from 372 artists curated from WikiArt. ArtSavant also recognizes if the identified style reappears in generated images. We leverage two complementary methods, including TagMatch, a novel inherently interpretable and attributable method, making it more suitable for broader use by non-technical stakeholders (artists, lawyers, judges, etc.). Our large-scale empirical study provides quantitative insight on the prevalence of artistic style copying across 3 popular text-to-image generative models. Notably, among a dataset of prolific artists, only 20% appear to have their styles at risk of copying via simple prompting of today’s popular text-to-image generative models. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper is about how AI-generated images can copy the style of famous artists without permission. The researchers wanted to understand why this happens and how to stop it. They used a new tool called ArtSavant, which helps identify an artist’s unique style by comparing their work to other similar paintings. The tool also checks if the same style appears in AI-generated images. The researchers found that only 20% of famous artists have their styles at risk of being copied without permission. |
Keywords
» Artificial intelligence » Classification » Diffusion » Prompting
