How to protect your art against AI using Glaze

Join us on social media!
It has been some time coming. AI took the art world by storm halfway into the year 2022. The biggest fear being it taking the jobs from artists. The biggest issue: It taking already existing art and art styles from artists in the blink of an eye, ignoring copyrights. The overall opinions have been mixed since then. Is AI taking work from artists? Well, yea to a degree. But individuals that really want to make a difference and put that little extra effort into their work don't see a decline in work. Does it steal art? Depending on who you ask the answer is either yes, or debatable. In the end, though, AI learning from copyrighted art and art styles is considered an issue among artists. So what can you as an artist do to minimize the risk of your art being stolen?

Hi! My name is Tessa, I’m a Dutch artist, art director, and creative project manager. I love to share my passion for this craft, nature, art and fantasy, and do that by creating this archive and community, alongside my company Tez Art & Design.

Table of Contents (Click to (un)fold)

What is AI

Your being here tells me that you probably already know what AI is, so I will keep it short. AI is short for Artificial Intelligence. It takes on many shapes and forms, programs like Google Assistant, Siri, ChatGP, and Alexa only being a few of them. In our art world, we see programs like Midjourney, Dall-E, Jasper Art, Starry AI, Nightcafé, Dream by Wombo, Synthesys X, Pixray, Deep Dream Generator, and the list goes on. Each and every one of them has a slightly different approach to AI learning, and many of them have different goals or target audiences.

AI, like any computer program that’s set up to be AI learns from human beings. Just like us by input from the outside world. This can be AI learning how to have a conversation with people, making assumptions about what you’re looking for on Google, but also how to make art. In doing so it will take from people and their behavior and actions what it’s programmed to learn from.

Unlike AI, we human beings can prove that we made an artwork ourselves. So any well-informed buyer of your art can confirm that it’s something made by you. Make sure you make use of these tips as they provide security for your client.

Why is AI even an issue?

Generally, this AI learning is no issue. It has been assisting us well through the past couple of decades or so. It does however become a problem when it starts learning from copyrighted things such as writing and drawings. In the writing world, it has been an issue for quite some time. Even when such programs are laid to bounds, there are still people out there exploiting its possibilities or creating new programs that can bypass the restrictions of others. This is not only causing copyrighted work to be stolen from people but also people mistakingly being accused of stealing their own work, sometimes causing them to lose their complete income because their selling account is being locked up.

The art world is a bit more fluid in that sense, we have more places to put out our work, which is a good thing of course. Just not if AI is grabbing everything it can from everywhere. It’s almost guaranteed that AI is using your art somewhere and your original artwork or style is still very recognizable within it. And that’s where AI is crossing the boundaries of what is okay and what is not. And yes, to a degree people using it are an issue too, that is, if they’re specifically using artist names to achieve a specific style or scene. Even more so if they try to sell it as their own. But that’s not the root of the issue. The roots lay somewhere you may not expect them to be.

So what is the root of the problem?

The fact that AI programs are open source programs. There is no one sole owner. This causes a whole heap of legal issues and I’m not the right person to explain them to you, but what it boils down to is that, because it’s open source, it can use anything it comes across on the internet without anyone ever being charged for breaking copyrights. And, if people just used AI as a reference and inspiration tool, there would be no real issue. But because people start selling it and the programs can’t be held accountable, there is no way to tame the beast.

What is Glaze

Glaze is a free tool that helps us artists to ‘cloak’ our art style. The program adds a filter known as a Style Cloak that overlays your artwork and turns it into a Cloaked Artwork so it’s hard for AI to distinguish what your art style actually is like and will think that this cloaked version is your style. This serves 2 main purposes. 1: It protects your style from being used by AI. 2: It will confuse AI and show a completely different style.

For example, when someone adds the prompt ‘Futuristic landscape + your name’ you only have to have a few landscape artworks online and AI can copy it in a matter of seconds. When Glaze is used, it makes very tiny changes to the artwork. For the human eye, this is barely visible, however, it can make AI ‘think’ that your style is similar to that of Van Gogh. It’s unable to distinguish your own unique style.

You can download it right here. Make sure that you pick the latest version for the best use.

Limitations of Glaze

There are some limitations to Glaze such as:

  • Glaze works best on textured artwork. It will work on other styles as well such as anime and other styles with flat colors, but it’s much more visible on styles like that. They are working on a solution for that problem though.
  • AI keeps evolving no matter what. And although Glaze will keep being developed, it may not always work as intended because of AI learning Glaze has no influence. So keep in mind that from time to time old cloaked art will become vulnerable. It might need to be updated with a new cloak.

Who has made this program?

Glaze is a free-to-use program. There is no profit coming from it and there is no political agenda. It’s made by an academic research group of Ph.D. students and CS professors that are interested in protecting internet users from invasive uses of machine learning. The team consists of:

  • Professor Heather Zheng
  • Faculty Lead Professor Ben Zhao
  • Collaborator Professor Rana Hanocka
  • Shawn Shan
  • Jenna Cryan
  • Emily Wenger
  • Karla Ortiz (artist)
  • Lyndsey Gallant (artist)
  • Nathan Fowkes (artist)
  • And 1100 test users


I highly recommend that you check out their website and download the program. I can copy everything they’re explaining right there. However, my goal is to make you aware. They explain everything step by step including visual examples. They continuously develop and improve Glaze and although it’s still a beta version: It’s extremely promising. I hope that like me, it gives you hope for a bright artistic future no matter what technology throws at you.

We would love to hear from you!

Sign in on Discord to comment and participate, or use the contact form.

Share on social media