A recent Vox article identifies several popular online platforms that are cashing in on the current AI lawsuits by selling their databases to LLMs. Companies that have cut deals include:
Facebook and Instagram are of course already compromised.
No wonder I can Glaze my artworks whatever I posted my artworks into Facebook and/or Instagram
The limitations of current Generative AIs can sooner or later make them goes the way of fidget spinners as those AIs need to stole things from datasheets to draw or write anything(unless when that AIs are True AIs), this aren't so bad if that generative AIs limit whatever they scrapes into their datasheets to public domain stuff
It's been a few years now of this and the AI still can't understand half the languages I speak, so it's arguably also anti-optimistic to hear about unless how this is going happens to be exaggerated. Discord comes to mind as well.
That probably means Threads, their twitter alternative, is also. Unfortunate, but not suprising.
Speaking of, I've heard before that Twitter is also scraping content for AI? It's a little confusing
I really should clone that repository that’s based off of Glaze on my laptop, oh yeah it’s Mist-project.
I’ll start using it for my DeviantArt posts, and on my artfol posts in the future, and a test drawing on here too.
Man I’m glad my gloomy sketch Twitter is private.
I honestly think this AI stuff that companies will destroy themselves over will end going the way of fidget spinners too, but who knows.
I know it's not what you mean, but I can attest to the last part. I was in an enormous server that discussed this at length. Every word you utter in VC and every two artworks posted in there is used to teach an AI somewhere to do something. If you value your copyright and your sovereignty, don't use VC or post art there. If you don't want to unknowingly contribute to the AI progression with who you are, don't do these things. If you If you oppose the idea of stolen AI learning, don't do those things. I would stop anyone from doing those things. I keep AI's fingers away from my input when possible at all times. People avoid it out of boycott and rightfully.
Aye, it's worth clarifying "scraping" is a general practice by bot crawlers (like search engine indexers and archive.org scripts), of which data mining for AI is one potential end purpose. I remember the Twitter spat and sudden censure of guest access, which I read as Elon coercing membership and/or confounding posterity by deliberately throttling reply chains amid the about-face on content policy. There are legitimate reasons to restrict bots, but as the article says, this is a self-serving bid to monopolize user data, not to protect privacy.
that's why I deleted my Threads account
I hardly use Facebook or Instagram anymore for this reason, and I never actually posted anything on Threads, I just kind of use it to see updates on SpongeBob, Waffles The Pug Mix, and a few other accounts I followed on IG in the earlier days. Meanwhile, I only used my current Tumblr page to post something to It Came from Tumblr, and repost cute animal stuff, I only granulated one of my pieces here on Side 7, just in case someone gets sneaky and goes under the radar about stealing stuff over here, since there's not really anything admins can do about off-site behavior, but I might feel forced to do that with my future art if this keeps going on all over the internet.
When I’m done with my personal site and booting it up and stuff, I also added a robots txt that will stop all kinds of image crawlers from all over the internet, and hopefully I’ll make my personal website/blog blacklisted on Google or something, I don’t know if I can do that.
Corollary to the OP, this Wired article provides an overview of company-specific user-end opt-outs from AI training. It also highlights Have I Been Trained?, a site to check and potentially counter-claim existing scrapes.
.png){kind=avatar}
