Just read that a huge number of digital art pieces are being used to train AI without the artists knowing

I was reading a report from a group called Spawning, and they said over 12 billion images from the web have been scraped for AI training. Many were taken from art sites without permission. I found this because I was looking into why my own style seemed to pop up in some AI generators. It makes me worry about posting new work. Has anyone else looked into protecting their art from this?

2 comments

2 Comments

dylan_thompson15h ago

Spawning's report on 12 billion scraped images is the tip of the iceberg. I saw my friend's portfolio, which is set to private on ArtStation, get mirrored on a random sketch site and then appear in a dataset. The tools to opt out, like "Do Not Train" tags, are a joke because the companies scraping ignore them unless you find every single copy of your work online. It's turned posting art into a security problem.

averywright13h ago

Check the terms on the sites you upload to, because that's where they get you. I mean, @dylan_thompson is right about mirrors, but the real issue is the fine print. A lot of big art platforms quietly added clauses that let them use your uploads for "research" or "AI development." So even if you avoid public scraping, you might have already agreed to it. I had to stop using one of my favorite portfolio sites because I finally read their update. Now I only host my own stuff, which is a pain, but at least I know where it is.