why would I use their AI? on top of that, wouldn't it be in their best interests to allow people to use their AI with as few restrictions as possible in order to maximize market saturation?

An important note here, the judge has already ruled in this case that "using Plaintiffs' works "to train specific LLMs [was] justified as a fair use" because "[t]he technology at issue was among the most transformative many of us will see in our lifetimes." during the summary judgement order.

The plaintiffs are not suing Anthropic for infringing on their copyright, the court has already ruled that it was so obvious that they could not succeed with that argument that it could be dismissed. Their only remaining claim is that Anthropic downloaded the books from piracy sites using bittorrent

This isn't about LLMs anymore, it's a standard "You downloaded something on Bittorrent and made a company mad"-type case that has been going on since Napster.

Also, the headline is incredibly misleading. It's ascribing feelings to an entire industry based on a common legal filing that is not by itself noteworthy. Unless you really care about legal technicalities, you can stop here.

The actual news, the new factual thing that happened, is that the Consumer Technology Association and the Computer and Communications Industry Association filed an Amicus Brief, in an appeal of an issue that Anthropic the court ruled against.

This is pretty normal legal filing about legal technicalities. This isn't really newsworthy outside of, maybe, some people in the legal profession who are bored.

The issue was class certification.

Three people sued Anthropic. Instead of just suing Anthropic on behalf of themselves, they moved to be certified as class. That is to say that they wanted to sue on behalf of a larger group of people, in this case a "Pirated Books Class" of authors whose books Anthropic downloaded from the book piracy websites.

The judge ruled they can represent the class, Anthropic appealed the ruling. During this appeal an industry group filed an Amicus brief with arguments supporting Anthropic's argument. This is not uncommon, The Onion famously filed an Amicus brief with the Supreme Court when they were about to rule on issues of parody. Like everything The Onion writes, it's a good piece of satire: https://www.supremecourt.gov/DocketPDF/22/22-293/242292/20221003125252896_35295545_1-22.10.03 - Novak-Parma - Onion Amicus Brief.pdf

My last comment was wrong, I've read through the filings of the case.

The judge has already ruled that training the LLMs using the books was so obviously fair use that it was dismissed in summary judgement (my bolds):

To summarize the analysis that now follows, the use of the books at issue to train Claude and its precursors was exceedingly transformative and was a fair use under Section 107 of the Copyright Act. The digitization of the books purchased in print form by Anthropic was also a fair use, but not for the same reason as applies to the training copies. Instead, it was a fair use because all Anthropic did was replace the print copies it had purchased for its central library with more convenient, space-saving, and searchable digital copies without adding new copies, creating new works, or redistributing existing copies. However, Anthropic had no entitlement to use pirated copies for its central library, and creating a permanent, general-purpose library was not itself a fair use excusing Anthropic's piracy.

The only issue remaining in this case is that they downloaded copyrighted material with bittorrent, the kind of lawsuits that have been going on since napster. They'll probably be required to pay for all 196,640 books that they priated and some other damages.

Copyright owners are then the new AI companies, and compared to now where open source AI is a possibility, it will never be, because only they will have enough content to train models. And without any competition, enshittification will go full speed ahead, meaning the chatbots you don't like will still be there, and now they will try to sell you stuff and you can't even choose a chatbot that doesn't want to upsell you.

they should have done that long ago, and if they haven't already started a backup in both europe and china, it's high time

This is the real concern. Copyright abuse has been rampant for a long time, and the only reason things like the Internet Archive are allowed to exist is because the copyright holders don't want to pick a fight they could potentially lose and lessen their hold on the IPs they're hoarding. The AI case is the perfect thing for them, because it's a very clear violation with a good amount of public support on their side, and winning will allow them to crack down even harder on all the things like the Internet Archive that should be fair use. AI is bad, but this fight won't benefit the public either way.

Anybody can use copyrighted works under fair use for research, more so if your LLM model is open source (I would say this fair use should only actually apply if your model is open source...).
You are wrong.

We don't need to break copyright rights that protect us from corporations in this case, or also incidentally protect open source and libre software.

The world you're envisioning would only have paid licenses, who's to say we can't have a "free for non commercial purposes" license style for it all?

The law is not the law.
I am the law.

insert awesome guitar riff here

Reference: https://youtu.be/Kl_sRb0uQ7A

I wouldn't even say AI is bad, i have currently Qwen 3 running on my own GPU giving me a course in RegEx and how to use it. It sometimes makes mistakes in the examples (we all know that chatbots are shit when it comes to the r's in strawberry), but i see it as "spot the error" type of training for me, and the instructions themself have been error free for now, since i do the lesson myself i can easily spot if something goes wrong.

AI crammed into everything because venture capitalists try to see what sticks is probably the main reason public opinion of chatbots is bad, and i don't condone that too, but the technology itself has uses and is an impressive accomplishment.

Same with image generation: i am shit at drawing, and i don't have the money to commission art if i want something specific, but i can generate what i want for myself.

If the copyright side wins, we all might lose the option to run imagegen and llms on our own hardware, there will never be an open-source llm, and resources that are important to us all will come even more under fire than they are already. Copyright holders will be the new AI companies, and without competition the enshittification will instantly start.

Well, theft has never been the best foundation for a business, has it?

History would suggest otherwise.

What you see as "spot the error" type training, another person sees as absolute fact that they internalize and use to make decisions that impact the world. The internet gave rise to the golden age of conspiracy theories, which is having a major impact on the worsening political climate, and it's because the average user isn't able to differentiate information from disinformation. AI chatbots giving people the answer they're looking for rather than the truth is only going to compound the issue.