Data as of Jul 25, 2026 · Based on 266 AI responses from ChatGPT Search and Google AI Mode · See how Parse measures this
For strongest protection, run 
Brands AI recommends here
Nightshade — best when you want to actively poison images to stop AI training: it inserts imperceptible pixel changes. Tradeoff: requires significant GPU memory and is offered free for non-commercial use.
Glaze — best when you want to protect artistic style from AI mimicry; apply Glaze after Nightshade for fuller protection. Caveat: a combined tool is planned but the two-step workflow is currently recommended.
WebGlaze — best when you lack high-end GPUs: a web/cloud version (invite-only) processes images for protection and is integrated into platforms like Cara for automatic protection on upload.
NightshadeSource | Citation share |
|---|---|
Nightshade: Protecting Copyrighthttps://nightshade.cs.uchicago.edu/whatis.html | 4.9% |
 Nightshade, the tool that ‘poisons’ data, gives artists a fighting chance against AI | TechCrunchhttps://techcrunch.com/2024/01/26/nightshade-the-tool-that-poisons-data-gives-artists-a-fighting-chance-against-ai/ | 3.7% |
 The AI lab waging a guerrilla war over exploitative AIhttps://www.technologyreview.com/2024/11/13/1106837/ai-data-posioning-nightshade-glaze-art-university-of-chicago-exploitation/ | 3.4% |
Glaze - Protecting Artists from Generative AIhttps://glaze.cs.uchicago.edu/ | 3.2% |
Downloads - Nightshadehttps://nightshade.cs.uchicago.edu/downloads.html | 2.6% |
 Nightshade, the free tool that 'poisons' AI models, is now ...https://venturebeat.com/ai/nightshade-the-free-tool-that-poisons-ai-models-is-now-available-for-artists-to-use | 2.1% |
Nightshade is here! AI data poisoning for artistshttps://www.aimeecozza.com/nightshade-is-here-ai-data-poisoning-for-artists/ | 1.8% |
 New Tech Gives Artists A Fighting Chance In The Battle Against Generative AIhttps://www.forbes.com/sites/robsalkowitz/2024/05/23/new-tech-gives-artists-a-fighting-chance-in-the-battle-against-generative-ai/ | 1.8% |
 University of Chicago researchers seek to “poison” AI art generators with Nightshade - Ars Technicahttps://arstechnica.com/information-technology/2023/10/university-of-chicago-researchers-seek-to-poison-ai-art-generators-with-nightshade/ | 1.8% |
 Log In ‹ The Transmission — WordPresshttps://www.unmc.edu/healthsecurity/transmission/2023/12/12/this-new-data-poisoning-tool-lets-artists-fight-back-against-generative-ai/ | 1.6% |
The main provider of Glaze and Nightshade is the Glaze Project, a research effort from the University of Chicago’s SAND Lab. They are not a commercial vendor; the tools are released free for artists and creators.
Glaze — created to protect an artist’s style from being learned or mimicked by generative AI image models. It applies subtle changes intended to make AI training systems perceive a different style while keeping the image visually similar to humans. glaze.cs.uchicago.edu
Official downloads:
Nightshade — a “data poisoning” tool intended to make scraped images less useful for unauthorized AI training by altering the data signals in the image. It is designed as a standalone tool and is aimed at discouraging model training on unlicensed artwork. nightshade.cs.uchicago.edu
Official downloads:
The project’s team includes University of Chicago researchers such as Ben Y. Zhao and Shawn Shan.
A few practical notes:
If you mean poisoning a broader personal data footprint (photos, text, websites, metadata, social profiles, etc.), Glaze/Nightshade only address a narrow part of that problem. I can outline other privacy-preserving approaches as well.
Both Nightshade and Glaze were developed by the SAND Lab (Security, Algorithms, Networking, and Data Laboratory) research team led by Professor Ben Zhao at the University of Chicago , alongside student researchers like Shawn Shan.
The project is run as an academic research and advocacy effort, supported by grants and donations (from organizations like the National Science Foundation, DARPA, and Amazon AWS) rather than a commercial enterprise. Their software is completely free for human artists to download and use.
You can safely access and download the official software directly from their dedicated university project pages:
Would you like assistance with how to set up and run Nightshade/Glaze on Windows or macOS, or do you need recommendations on best practices for posting poisoned images online?
The main source for both tools is the University of Chicago SAND Lab / The Glaze Project. They are free tools rather than commercial products.
Glaze — designed to disrupt AI models' ability to learn and mimic an artist's visual style. The current official download page lists Glaze 2.2 for Windows and Glaze 2.1 for macOS.
Nightshade — designed to create "poisoned" training samples intended to disrupt models trained on scraped images. The official site currently lists Nightshade 1.1 and provides its standalone download and user guide.
WebGlaze — browser-based Glaze for people without suitable desktop hardware. It's currently invite-only and sends uploaded images to University of Chicago's processing infrastructure, unlike the standalone Glaze app, which is designed to run locally.
One important distinction: Glaze is primarily defensive/style protection; Nightshade is the poisoning component. The researchers describe Nightshade as optional and warn that its effectiveness isn't guaranteed against future countermeasures.
If by “public data” you mean photos, illustrations, or other images you're publishing online, these are the official tools I'd use rather than downloading similarly named third-party versions.
Nightshade and Glaze are offered exclusively by The Glaze Project , a research initiative run by computer science professors and PhD students at the SAND Lab (Security, Algorithms, Networking, and Data) at the University of Chicago , led by Professor Ben Zhao and researcher Shawn Shan.
The tools are free-to-use, non-profit academic research releases funded by grants (such as the National Science Foundation, DARPA, and C3.ai).
Are you looking to run these tools on a local desktop app (using a CPU or GPU) or trying to access WebGlaze , and do you need help with the system requirements?
Nightshade and Glaze are not commercial products offered by a private tech company; they were developed and are maintained by a research team at the University of Chicago (specifically the SAND Lab led by Professor Ben Zhao).
The tools are completely free for artists and individual creators to use. You can access and download them directly through official academic project channels:
@TheGlazeProject on Instagram or X/Twitter).If you want, let me know:
I can help you figure out if you should run the local tools or apply for a WebGlaze cloud invite.
The main provider is the Glaze Project at the University of Chicago’s SAND Lab, led by Ben Zhao and colleagues. They develop both Glaze and Nightshade specifically to help creators protect images from unauthorized AI training.
Official source/download:
One important distinction: these are primarily image-protection tools, not general-purpose anti-scraping systems. They don't prevent someone from downloading or copying your public data; they attempt to make images less useful for particular kinds of AI training. The project's own research describes Nightshade as a poisoning attack against text-to-image training.
If by “public data” you mean text, photographs, PDFs, audio, or an entire website, I can also give you the current landscape of Nightshade/Glaze-like defenses for each type.
Nightshade and Glaze are not commercial products offered by a private tech company; they were developed and are maintained as free, open academic research projects by the SAND Lab (Security, Algorithms, Networking, and Data) at the University of Chicago , led by Professor Ben Zhao.
The tools are provided directly to creatives at no cost through their official project websites:
Key details regarding their distribution and usage:
Are you planning to run these tools on a machine with a dedicated GPU (like an Nvidia card) or via CPU mode , and do you need help understanding how they differ in processing time?
The main source for both is the University of Chicago SAND Lab / The Glaze Project—they aren't commercial products from a third-party vendor.
One important distinction: Glaze is a style-protection/cloaking tool, while Nightshade is specifically the data-poisoning tool. The researchers themselves describe Nightshade as a last-resort defense against unauthorized scraping/training.
If by “public data” you mean photos, illustrations, or other images you've posted online, I can also compare Nightshade vs. Glaze and explain what each actually protects against and their current limitations.
The primary source of both Glaze and Nightshade is the University of Chicago's SAND Lab (Security, Algorithms, Networking and Data Lab). They are research tools, not commercial products, and are offered free of charge by the Glaze Project.
You can get them from the official project:
The project includes:
It's also worth noting that these tools are an active area of research. The Glaze team has continued releasing updates intended to improve robustness, while independent researchers and practitioners have debated how effective these methods remain against newer training pipelines and possible countermeasures. In practice, effectiveness may vary depending on how images are collected, processed, and used by different model developers.
Nightshade and Glaze are developed and offered by researchers at the SAND Lab (Security, Algorithms, Networking, and Data) at the University of Chicago , led by professors Ben Zhao and Heather Zheng along with their student research team (operating as The Glaze Project).
The tools are completely free to use, non-commercial, and designed to protect human creatives from unauthorized AI scraping and style mimicry.
You can access and download the official software directly from their dedicated project websites:
Are you planning to run the tools locally using a GPU (Windows/Mac) or looking into the WebGlaze invite process? I can share system requirements or best practices for processing your images.
