ThatPainter is reader-supported. When you buy through links on our site, we may earn an affiliate commission. Learn More
Nightshade is a free research tool from the University of Chicago’s Glaze Project that alters images to create “poison” samples. Its intended effect depends on those images and their descriptions being included in a model’s training data: the samples may then shift how the model associates visual features with targeted text concepts. The researchers have demonstrated effects in experiments, but that is not a guarantee that Nightshade will affect every AI generator or protect every artwork.
What Nightshade does
Nightshade is designed to give artists a technical response to the use of their images in AI training without consent. The software makes changes intended to be hard for people to notice but meaningful to a model learning relationships between images and text. If a treated image and its description enter a training set, the sample may pull a targeted concept toward an incorrect or unexpected association.
The Glaze Project describes an example in which a prompt for a cow flying in space could produce an unexpected object after training on poisoned images. That illustrates the intended mechanism, not a predictable outcome for every prompt or model. Nightshade cannot ensure an image will be scraped or used, and its effect depends on what happens during training.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
#1 Best Overall
What the research found—and what it does not prove
The Nightshade paper reports experiments against specified research models. Its authors say optimized attacks could corrupt targeted prompts in SDXL with fewer than 100 poison samples under the paper’s experimental conditions. The paper also reports high attack success with larger sample counts in its tested setup. Those findings describe controlled experiments; they do not establish that a named commercial generator has been affected or that the same result will occur in a deployed service.
The Glaze Project reports more than 2.5 million Nightshade downloads since January 2024. That is the project’s cumulative figure, not an independently audited count, and downloads do not show how many images were used in training or what impact they had.
Nightshade and Glaze have different jobs
| Tool | Stated purpose | How it works at a high level |
|---|---|---|
| Nightshade | Discourage training on images without consent | Uses prompt-specific poisoning intended to affect learned image-text associations if treated images enter training data. |
| Glaze | Disrupt AI style mimicry | Uses style cloaking intended to make artwork appear stylistically different to AI models while remaining similar to human viewers. |
They are related tools from the same project, but they are not interchangeable: Nightshade targets training associations, while Glaze is aimed at style mimicry. The Glaze Project says both tools are free for artists and will not be used to generate profit.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What Nightshade cannot do
- It does not remove work from a model that has already trained on it. Nightshade’s proposed mechanism acts through training data; the cited sources do not establish that it reverses previous training.
- It is not a guarantee of protection. The image must enter training for the proposed effect to matter, and the experimental results do not prove a universal effect across commercial systems.
- It is not a legal remedy. The project presents Nightshade as a technical response, not a replacement for copyright law, licensing, or legal advice.
University of Chicago computer scientist Ben Y. Zhao described the motivation this way: “What we needed was something technical to push back, and that’s what Nightshade is for, is to give a little bit of teeth to copyright in the wild and allow content owners in general to resist this idea that whatever you put online is anyone else’s training fodder.”
Quick Recap
Sources
- The Glaze Project: Nightshade and project information
- Nightshade research paper and abstract
- University of Chicago News: Big Brains podcast
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
