Why would anyone put anything of value on the internet at this point?
Sometimes, unfortunately, you have to put your foot down and never be nice to those filthy bastards by carefully choosing who gets to be in your platform. Start networking with the likes of EFF and other creatives’ groups opposing AI slop, attend next year’s DefCon to call attention to your plight and gain help to protect your art.
NGL, I saw the writing on the wall decades ago with deviantart and it’s why I pulled down pretty much everything I used to have up. The only real option other than a private forum is using old school listservs or e-groups where vetted users post and recieve artworks via an email system.
I’ve beat this drum before, but this has the same gotcha as permissively licensed code.
You put it out there, and people WILL do things with your code you don’t like. And you cant really stop it if they are motivated.
In absence of legal protection (which scrapers dont seem to care about these days), you pick exposure, or control, on a spectrum. You cant have your cake and eat it.
I find it extremely sad in the Cara situation, because this person is basically getting trolled, and by their own words they are already close to burning out.
The people scraping Cara are not AI companies looking for cheap training data. They are just twitter trolls toying with someone. They are not gonna spend the money and expertise required to use that art to train an AI, at best they’ll make a shitty LoRa which nobody will use because it will have no value. They are not doing it for any nefarious reason other than : 1. it’s cheap, 2. it works (the admin responds with intensity), 3. it is not illegal.
So the Cara admin is gonna compound their own burnout by piloting a crowdfunding campaign that will end up in heartbreak because no honest lawyer will take a case like that. And this will do nothing to advance Cara’s original mission which is to be a social space free from AI slop.
I wouldn’t be the least surprised if Cara ends up closing shop, all because the owner wasn’t clear-headed enough to disengage from internet trolls.
The people scraping Cara are not AI companies looking for cheap training data.
How are you so certain about this?
Could have put a link to the fund in question. https://www.gofundme.com/f/help-cara-legal-fund
I assume this is the correct one, but I dont know as I am not on that platform, and cant get it first hand.
https://www.gofundme.com/f/help-cara-legal-fund/cl/d?utm_campaign=pd_ss_icons&utm_content=amp30-treatment-4&utm_medium=customer&utm_source=copy_link&lang=en_US
This is what I found
FYI, both links are the same, except that yours is filled with random trackers. As a rule of thumb, basically everything past a question mark can be safely deleted off a link adress to clean it from unwanted trackers.
Oh ok I didn’t know that, thanks for letting me know
You can also get the firefox extension ClearURLs which will do it for you automatically every time you open a link, and they have a little box where you can paste any URL and it will clean it up and give you a clear one to share.
I’ve been using the duck browser, but I’ve been meaning to give Mozilla a try again. I’ll look into it, thanks!
It also works with forks, like waterfox or librewolf
Don’t they use Anubis?
i just visited and no, they don’t seem to use it, or at least i could access the site without it. but even if they did, anubis is incapable of defending against targeted scraping.
Anubis can do an okay job… (at least so the whole site isn’t being actively DDos’ed) at the cost of compromising user experience a lot by sending harder challenges more often. It kinda sucks that the average user might not have that fast of a CPU, while motivated scrapers might spend tonnes on compute just out of spite.
(i am a bit biased on this though)
anubis does its thing by making you waste your cpu cycles and then giving you a cookie, if you use that cookie for subsequent requests then you’re free to browse the site. it might in theory slow down dumb scrapers that don’t save that state and parallelize the mass scraping via multiple ip addresses (since without the cookie they’ll have to solve the challenge for each request) but if one’s making a targeted attack for one specific site, they don’t do any of that. it should be trivial for a bad actor to solve the challenge once and use the cookie, like a regular user would. anubis only slows down scrapers that scrape all sites indiscriminately and don’t expect it, once you know it’s there then it’s easy to bypass (and if you can afford to scrape the entire site you can surely afford to get past anubis once).
Have they tried contacting the EFF?
The more time goes on, the more I realize Jay and Silent Bob at the end of Jay and Silent Bob Strike Back were correct.
Anyone up for invite only art networks?
I think it’s the only way
Money would be better spent on a hitman.
I don’t think it’s going to stop until like armed people show up at the scraper’s homes. But I doubt the state would send the police for something like this, because there’s not enough money at stake.
For what exactly? Web scraping publicly available data is not a crime in itself.
Look up Aaron Swartz
See previous comment about not enough money at stake.
Who or what is Cara?
Anti-AI art platform intended for artists to host their portfolios.
You can poison your art against web-scrapers now, it’ll fuck up any AI models that try to use your content, I think there’s multiple ways to do it but the one made by the university of Chicago is called Nightshade: https://nightshade.cs.uchicago.edu/whatis.html
AFAIK Cara has the option to poison images using glaze.
This is working less and less because there’s also tools to un-poison that content before ingestion. Scrapers are also getting good at detecting poisoned content before ingestion as well.
Welcome to the arms race.
💚
Is it publicly accessible? Then it is not practical to defend.
If a customer can view your art on their device than they can save it. Same deal as with streaming videos. Piracy cuts both ways.
Piracy cuts both ways.
how are the artists engaging in any piracy?
what an astoundingly obtuse take
It is very different legally if you are getting economic benefits from scrapping. If it was done by data brokers to sell the images or an AI company to train a model then you can definitely sue them.
Whether the scraper seeks profit or not doesnt change the technical issue.
…And if the for-profit scrapers were scared of legal trouble, they wouldnt be doing it.
and yet if you pay for Netflix and torrent black mirror withour a VPN you will get a cease and desist… curious!
/hj
Most people on here are pro piracy. And consequences for piracy are rare.
The distinction:
Piracy is an availability problem.
Taking someone else’s art and passing it off as your own is taking credit for someone else’s work and severs all connections between artist and consumer.
Someone else’s work not being available at a “competitive” price for training is a problem here.
An apple tv exclusive being pirated by someone without apple tv is very similar.
Using someone else’s work to create a competitor is not piracy
Sure is. You are just making a copy without compensation and using it as you see fit.
Still no.
Theyre rare if you know what youre doing, sure.
They’re rare generally. C&Ds for torrenting in many countries are practically non-existent, unless you’re a big uploader.
Yes.
Thats basically a truism, and it has nothing to do with the scraper’s motivations or legality.
I dont understand why you are getting downvoted.
Maybe I can shed some light here, but maybe it’s also just arguing on the internet… The problem is the following: imagine you’re an artist with a physical shop. People go to your shop, they like your art and take pictures of it with their phone. That’s the second best thing that can happen to you. (after someone buying your art or hiring you for commissioned work ofc). Bc the artist gets popularity from it. Now imagine somebody sets up a new storefront in front of the artists shop. Shows the artists work, but takes away the name. Says they can commission any piece that’s like the artist’s work, but for free. The artist makes no revenue anymore. They don’t get popularity, bc the new shop shows it’s own name and hides the artist. Now the artist goes bankrupt. The fake store front moves over to the next artist’s shop and repeats the trick. Until they’re all out of work.
Many artists like piracy. Bc it means people talk about your work. It gets shared and makes you popular, which inevitably leads to more buying customers at some point. Ai scrapers though mean you’re not gaining in popularity anymore. So you starve, and that’s the problem
We’re not talking about AI scraping generally, we’re talking about a space that tries to replicate something between Instagram and DeviantArt while keeping scrapers out.
I say that’s impossible to sustain, unless they make the site highly inaccessible.
I’m not saying I don’t want it to; that’d be great if it could. But they are fighting an impossible battle, IMO.
People on here hate AI and like piracy/not paying for content. Makes perfect sense to me. Doesn’t change my mind. They want to dunk, at least that is what it feels like.
