Please share your thoughts about this as well as spread the word!
Thanks so much. ^_^ ♥
Transcript:
Sometime around June 3, 2024, Instagram and FB announced that it has plans to use people’s public posts and images to help train AI tools across meta services. This includes both original artwork and regular photos (i.e. selfies, family pics, etc.) which will be used to train the company’s AI image generator.
This is due to go into effect on Wednesday, June 26, 2024. Any posts, images, image captions, comments, and stories that are shared publicly will be used as part of training Meta AI. It has also claimed that it will use data dating all the way back to 2007. However, it will not include anything sent through private messaging.
While it is possible to opt out of posts being used for meta ai training in Europe iirc due to EU laws, IG & FB have made it nearly impossible to opt out in the US.
Some artists have decided to leave Instagram altogether and move to Cara. Cara, founded by Jingna Zhang, is a portfolio app that protects artists’ work from being used as training gruel for AI. AI content can be posted only if it’s clearly labeled. Listed on Cara’s FAQ page, the app protects its users from AI scraping by implementing “NoAI” tags on all of its posts. The website says these labels “are intended to tell AI scrapers not to scrape from Cara.” Although Cara’s labels aren’t a “fully comprehensive solution and won’t completely prevent dedicated scrapers”, it is a step towards rejecting complacency to AI stealing the work of artists, without their consent.
Unfortunately, this doesn’t seem to stop at just Instagram and FB. Right now, almost everything posted publicly on the internet is considered fair game for AI training. It's going to become more and more important to carefully read all terms and conditions in user agreements to ensure there aren't any hidden statements giving the company access to your data for AI training purposes.
According to an article from: College of Saint Benidict Saint John’s University (https://guides.csbsju.edu/AI-Images#:~:text=%22To%20create%20AI%2Dgenerated%20images,image%20and%20text%20fit%20together) even sites such as Canva use AI. And their image sources are not restricted to just Instagram or FB. "To create AI-generated images, the machine learning model scans millions of images across the internet along with the text associated with them. The algorithms spot trends in the images and text and eventually begin to guess which image and text fit together.”
With this knowledge, it really doesn’t feel like anywhere is completely safe to post your art or images anymore.
Can anything protect art?
Image cloaking. As described by the website gelato. https://www.gelato.com/blog/how-to-protect-your-art-from-ai
Cloaking images with Glaze or Nightshade
Have you ever wanted to use a tool that shields your artwork from AI, making it invisible (well, almost) to their prying algorithms? This is exactly where image-cloaking tools like Glaze and Nightshade come in, offering a powerful way to deter AI from using your artwork.
How does image cloaking work?
Both Glaze and Nightshade employ a technique called "adversarial cloaking." They analyze artists' artwork and introduce subtle, imperceptible changes to the pixels. These modifications don't affect the human eye's perception of the image, but they confuse AI algorithms trained on standard datasets. Think of it like a visual camouflage, making your artwork appear different from AI while remaining unchanged for you.
Deters AI training: Cloaking your artwork makes it less effective for AI model training, which discourages AI from replicating your unique style.
Promotes originality: With AI adept at mimicking styles, cloaking adds a layer of protection. It hinders AI from easily capturing the essence of your artwork.
Offers peace of mind: Knowing your work is less susceptible to AI imitation can be a significant source of comfort.
However, there are a few important factors you must keep in mind:
It's not foolproof: While effective, cloaking isn't an impenetrable shield. Highly advanced AI models might still be able to mimic your style to some degree.
Software availability: Currently, both Glaze and Nightshade are under development. While Glaze is available for download, Nightshade's release is pending.
Potential ethical concerns: Nightshade specifically might introduce noise into AI training datasets, potentially hindering overall AI development. This raises ethical considerations that are still being debated.
As an artist – Overall, in the end it really seems that aside from programs to shield your art, there seems to be no escape from AI learning from your images.
Leaving the only three options as: shield your art before you post it *anywhere.* Don’t upload it at all, or just upload as is and hope for the best.
As an artist supporter – As a supporter, now more than ever we really need to check with the artist *before* sharing their art. Make sure that the artist is ok with it. If an artist has taken down their images, but you share it on your platform anyway, then their art still can and will be used by Meta AI. It’s the equivalent of you literally stealing the art from the artist yourself and handing it over to AI to do with it as it pleases. And that’s just rude. Please do *not* do that. Ask first!
(The only other acceptable alternative to asking is if the artist has already posted in their BIO or on their page somewhere that it is in fact, ok to share their art *with the proper credit* being given to them.)
For your personal account – Really, the option here is the same as it is for artists as it is for our personal family photos and selfies. Either shield it first before you upload it, don’t upload it at all, or just upload it anyway and hope that you never see you or your family’s likeness on anything you’re appalled by.
FOR ARTISTS TO ANSWER – WHAT NOW?
Now armed with this knowledge….
1 - How will you personally be sharing your art in the future?
2 – Are you ok with others sharing your art? Why or why not?
3 – What’s the best way for your fans to support you? ^_^