| Submit your comments on this article |
| Economy |
| AI industry horrified to face largest copyright class action ever certified |
| 2025-08-11 |
| [ArsTechnica] Copyright class actions could financially ruin AI industry, trade groups say. AI industry groups are urging an appeals court to block what they say is the largest copyright class action ever certified. They've warned that a single lawsuit raised by three authors over Anthropic's AI training now threatens to "financially ruin" the entire AI industry if up to 7 million claimants end up joining the litigation and forcing a settlement. Last week, Anthropic petitioned to appeal the class certification, urging the court to weigh questions that the district court judge, William Alsup, seemingly did not. Alsup allegedly failed to conduct a "rigorous analysis" of the potential class and instead based his judgment on his "50 years" of experience, Anthropic said. If the appeals court denies the petition, Anthropic argued, the emerging company may be doomed. As Anthropic argued, it now "faces hundreds of billions of dollars in potential damages liability at trial in four months" based on a class certification rushed at "warp speed" that involves "up to seven million potential claimants, whose works span a century of publishing history," each possibly triggering a $150,000 fine. Confronted with such extreme potential damages, Anthropic may lose its rights to raise valid defenses of its AI training, deciding it would be more prudent to settle, the company argued. And that could set an alarming precedent, considering all the other lawsuits generative AI (GenAI) companies face over training on copyrighted materials, Anthropic argued. |
| Posted by:Grom the Affective |
| #14 Structurally, isn't that fundamentally what colleges do or suppose to do with students? Pretty much exactly that! Take an art student (please!). In Art History, they look at every famous painting ever done. The images influence their brain, but there is no place where copies of the Mona Lisa or Starry Night are stored. Looking at the images just changes the interconnections in the student's brain. A neural network works the same way. It has to be trained. The data it sees influences the interconnections of the network, but nowhere is a copy of the image saved. If you borrow a book from a library, reading it is not a copyright violation. What about the AI 'hallucination answers' that almost always crop up these days? Modern AI comes in two flavors: classifiers that distinguish this from that, and generators that produce new stuff. Both types are based on a model inspired by the human brain - the neural network. A neural network consists of multiple layers of "neurons stacked together. Each neuron in one layer is connected to every neuron in the next layer. Training the network changes the strength of the interconnections between layers. All very brain-like. When people say "AI" nowadays, they are usually referring to generators like ChatGPT and the various text-to-image programs. The main difference between generators and classifiers is that classifiers do useful things like recognizing hand-written digits or reading hospital x-rays. Generators produced stochastically generated slop that may look pretty or sound meaningful, but all the AI is doing is rolling dice. When chatbot like ChatGPT produces silly or factually incorrect output, we say it hallucinates. I would argue that *everything* a chatbot produces is a hallucination. Some of them are simply more plausible than others. If you ask, "What is a horse?", you might get "A horse is a large mammalian quadruped of the genus Equus. Or you might get "A horse is a horse, of course, of course. And no one can talk to a horse." We would call one an answer and the other a hallucination, but from the AI point of view, both are produced by exactly the same process and the AI is unable to distinguish between them as far as truthyness. A chatbot (formally a Large Language Model) has no internal concept of Truth. It is not a database of facts. All it knows about are tokens - words, phrases, characters and punctuation - and how frequently one token follows another in the training data. Your chatbot may generate a nicely formatted legal brief, but the legal case references may be actual legal decisions or they might be made-up AI slop. The chatbot neither knows or cares. It is hallucinations all the way down. tl;dr: Anyone asking ChatGPT serious questions has failed the Turing Test. |
| Posted by: SteveS 2025-08-11 21:27 |
| #13 Elon Musk’s AI Chatbot Grok Briefly Suspended from X After Going ‘Unhinged’ |
| Posted by: Skidmark 2025-08-11 17:59 |
| #12 But it would be wise and considerate if AI doesn't copy those other sources word for word What does it think it is? another future Donk politician or nominee to the Federal bench? |
| Posted by: Procopius2k 2025-08-11 17:05 |
| #11 Don't copy that floppy! |
| Posted by: Claith Speash6417 2025-08-11 16:23 |
| #10 “ChatGPT, Am I The Kwizach Haderach?” |
| Posted by: Grom the Affective 2025-08-11 14:09 |
| #9 Who were the lawyers suing Fred for re-posting articles? They got a slice but then got shutdown. |
| Posted by: Skidmark 2025-08-11 13:54 |
| #8 #4, Elmert, I think we went through something like this when Apple/Steve Jobs sued Microsoft/Bill Gates over Microsoft's graphical user interface (GUI) claiming that Apple developed the GUI first. IIRC the final ruling was that Microsoft could legally develop a GUI with a similar look and feel as Apple's as long as they did not copy Apple's source code. Fact is, Jobs got the idea for a GUI from Xerox's Palo Alto Research Center (PARC). It seems the suits at Xerox did not understand or appreciate what had been developed at PARC and let it wither on the vine. AI might be able to summarize information from other sources. But it would be wise and considerate if AI doesn't copy those other sources word for word, image file for image file and audio file for audio file without even crediting and providing links to the sources. I'm sure there's more to this story than what I've read here but it sounds like AI is arguing it is "too big to fail" and therefor exempt from copyright laws. |
| Posted by: Abu Uluque 2025-08-11 13:17 |
| #7 What about the AI 'hallucination answers' that almost always crop up these days? Evidence of genuine AI creativity. |
| Posted by: Elmerert Hupens2660 2025-08-11 12:49 |
| #6 The US is taking a cut from chip sales to China - what does it mean? |
| Posted by: Skidmark 2025-08-11 09:17 |
| #5 What about the AI 'hallucination answers' that almost always crop up these days? |
| Posted by: Mullah Richard 2025-08-11 09:17 |
| #4 Has there been any creator of copyrightable works in the last millennia who was not trained, formally or informally using other copyrightable works? By this logic any creator would have to prove that he had never been exposed to any work that is even only remotely related to his own work in a rough analogue to clean room design of reverse engineered software. E.g. James Cameron would have to pay creators of copyrighted non fiction reports and documentaries on the Titanic disaster unless he could prove that no one involved in the film's development had ever been exposed to such works. |
| Posted by: Elmerert Hupens2660 2025-08-11 08:54 |
| #3 A college/university education is supposed to teach the student how to think — wisdom, if you will. Knowledge should be picked up along the way, just as part of the process that should continue life long. My favourite essay on the subject: Examsmanship and the Liberal Arts: A Study in Educational Epistemology |
| Posted by: trailing wife 2025-08-11 08:26 |
| #2 I believe this is refered to as 'the tragedy of the commons'. |
| Posted by: ed in texas 2025-08-11 08:10 |
| #1 Structurally, isn't that fundamentally what colleges do or suppose to do with students? Cram a lot of information in the mush of the brain. |
| Posted by: Procopius2k 2025-08-11 06:32 |