about
Adversarial Perturbations Cannot Reliably Protect Artists from Generative AI (arxiv.org)
5 points by dpaleka on Jun 19, 2024 | hide | past | pdf | discuss on HN

In plain words: Popular tools add tiny invisible changes to artwork so AI image generators can't copy an artist's style. But simple fixes like upscaling wipe out those changes, and a user study found every protection was easily bypassed.

Abstract · Adversarial Perturbations Cannot Reliably Protect Artists From Generative AI

Artists are increasingly concerned about advancements in image generation models that can closely replicate their unique artistic styles. In response, several protection tools against style mimicry have been developed that incorporate small adversarial perturbations into artworks published online. In this work, we evaluate the effectiveness of popular protections -- with millions of downloads -- and show they only provide a false sense of security. We find that low-effort and "off-the-shelf" techniques, such as image upscaling, are sufficient to create robust mimicry methods that significantly degrade existing protections. Through a user study, we demonstrate that all existing protections can be easily bypassed, leaving artists vulnerable to style mimicry. We caution that tools based on adversarial perturbations cannot reliably protect artists from the misuse of generative AI, and urge the development of alternative non-technological solutions.

Robert Hönig, Javier Rando, Nicholas Carlini, Florian Tramèr
arXiv:2406.12027 · cs.CR · submitted Jun 17, 2024 · updated Feb 11, 2025
abstract · pdf · html

add comment on HN
Also discussed: Jun 2024 (4 points, 1 comment) · Jun 2024 (2 points, 0 comments)