Ask Lemmy

If you're mobile-only, that's probably going to be a problem.

The only tools I know are Glaze and Nightshade. Think of them as defense and offense, respectively. They both introduce subtle changes to your image that make it harder for AI to use it. Well I say subtle - the stronger the protection the more noticeable the effect is to the human eye.

Glaze is defense, meaning AI can't use the image to imitate your style. I don't personally use it because I'm simply not well-known enough for anyone to want to specifically imitate my nonsense.

Nightshade is offense, it poisons the AI's data with useless information. Basically it makes the AI think that an image of, say, a cat is a chimney. That's what I use on anything I upload.

You can use both, though I don't remember if there's a recommended order in which to use them. The effect will then also be more noticeable. If you decide to use one or both, you should play around with the settings to see what's tolerable to you and what makes sense for an individual image.

Last time I looked, they were working on a mobile solution but I haven't heard anything on that front, unfortunately.

Post them privately, or not at all.
Especially if you want to sell anything. Then only post your art in that store.

Others have already mentioned the tools.

This is a losing battle. The systems will continue to improve, obfuscation will always lag behind. My advice would be to learn to embrace it. Zen meditation can help.

Imaginably, this response may be very annoying; seeing the current trajectory of development, it is hard to imagine you can. Not sharing is the only defeat, which also defeats the purpose of art. Quite a conundrum.

Not an expert on this topic but I've read about it a fair bit and tinkered around with image generators:

You don't post them, basically. Unfortunately nothing else will really work in the long term.

There are various tools -- Glaze is the first one I can think of -- that try to subtly modify the pixels in the image in a way that is imperceptible to humans but causes the computer vision part of image generator AIs (the part that, during the training process, looks at an image and produces a text description of what is in it) to freak out and become unable to understand what is in the image. This is known as an adversarial attack in the literature.

The intention of these tools is to make it harder to use the images for training AI models, but there are several caveats:

• Though they try to be visually undetectable to humans, they can still create obviously visible artifacts, especially on higher strength levels. This is especially noticeable on hand-drawn illustrations, less so on photographs.
• Lower strength levels with fewer artifacts are less effective.
• They can only target existing models, and even then won't be equally effective against all of them.
• There are ways of mitigating or removing the effect, and it will likely not work on future AI models (preventing adversarial attacks is a major research interest in the field).

So the main thing you gain from using these is that it becomes harder for people to use your art for style transfer/fine-tuning purposes to copy your specific art style right now. The protection has an inherent time limit in it because it relies on a flaw in the AI models, which will be fixed in the future. Other abusable flaws will almost certainly remain and be discovered after the ones currently used are fixed, but the art you release now obviously cannot be protected by techniques that do not yet exist. It will be a cat-and-mouse game, and one where the protection systems play the role of the cat.

Anyway, if you want to try it, you can find the aforementioned Glaze at https://glaze.cs.uchicago.edu/. You may want to read one of their recent updates, which discusses at greater length the specific issue I bring up here, i.e. the AI models overcoming the adversarial attack and rendering the protection ineffective, and how they updated the protection to mitigate this: https://glaze.cs.uchicago.edu/update21.html

I know about this tool, but haven't used it myself

https://glaze.cs.uchicago.edu/

Glaze is a system designed to protect human artists by disrupting style mimicry. At a high level, Glaze works by understanding the AI models that are training on human art, and using machine learning algorithms, computing a set of minimal changes to artworks, such that it appears unchanged to human eyes, but appears to AI models like a dramatically different art style.

It depends on what exactly you want to achieve. If you want money, then upload it to Adobe, Shutterstock, and such. That's the best offer you can expect. No one will offer you big bucks for an AI license just because you have protected your drawings.

If you want to have your own site, then you could rely on Cloudflare for handling the technical and legal side of preventing scraping/AI use.

But I guess the main worry for any artist is other artists who use AI. That's where Glaze and Nightshade come in. It's already been suggested but you should know how much you can expect.

These tools target the original Stable Diffusion 1.5. IIRC they also work on SD 2.0 because they reused some components. I am not sure what other versions, if any, could be affected. Certainly not the newer ones.

It goes without saying that the major companies were never affected, could not be affected. Since no one mentions it, I guess it's self-evident but I want to repeat it for the uninitiated.

I think these early models are still used partly because they have lower hardware demands and partly because they are less professionally censored (ie more suitable for porn).

Anyway, the effectiveness against hobbyists, your competitors, or other small scale AI users, is also limited. They may not use a susceptible model, especially if they make SFW images. If their model is susceptible, then these tools may waste a few hours of their time and maybe a bit of money. But it won't get rid of the competition or even significantly harm them.

I have an idea: if every artist/photographer etc uploaded their work with a watermark across it that said "f*ck you" then the LLMs that steal it would all start telling their users "f*ck you" as well.