A lot of knowledge in such niches has never been written down. Papers don't contain it. Tacit knowledge, it's in unpublished config files and unpublished experiments that some prof or PhD student remembers, all the things that are not "publishing-worthy" (because not flashy, not novel-sounding enough, or a negative result etc) but crucial for the method to actually work. Most of expertise is like this and this is why spending years in a research group, talking to the researchers, participating in their culture, understanding their thinking over years (or even just months) can be invaluable. There are also many things that people only say after a beer, after you know each other well. Claude and ChatGPT haven't ingested any of this. They ingested the sanitized surface version, the press release / advertisement version. Because papers are basically career marketing and advertisement mostly. The real expertise is guarded and hidden.
This is of course about the PhD level and beyond. For regular university-level education, where the goal is to pass the usual exams, ChatGPT and Claude are more than enough. But for real-world "doing", not (yet).
That's not unique to tech startups. Almost nothing in terms of explicit methodology has any effect on education outcomes. The literature on education research is of abysmal quality too.
There is not a huge market of people who actually want to learn something on their own motivation. Get some piece of paper, yes. So the huge market will only materialize if the degree granting institutions adopt something like this.
Why did you sign off your copyright to ACM then? If you kept the copyright and prevented others from distributing your work, you could be the sole distributor and negotiate the price with the AI company yourself.
Ok, then next question is why are you in academia? To share knowledge with humanity, bring the Promethean light to the mortals or something I guess. But then if they use it for productive things that make money, that's too smelly, that's too practical, too dirty, or what?
What a petty view of others you have :( What should they do then? Just shrug and move on, the machine must continue to grind no matter what? Why there is always this request for 100% purity if you dare to criticize the system or part of it?
that was a big jump from what I said. and what I said was that research lives off grants so researcher can't really come back and cry about strings attached to grants. ofc the problem is faceted, as grant go to published authors with track record, creating a publishing mafia, but everything rotates around the cashflow.
but you want sweet grant money, you play the game by the rules.
researcher that don't want to play the grant game can go and find a VC to subsidy product creation
If you don't hold a patent for the use of the knowledge you published publicly, you can't prevent others from using the knowledge. You enjoy the prestige attached to the idea that you're an academic who participates in giving away their knowledge but then you play this game when that knowledge would actually be useful as opposed to being read by 3 other people in your special area who sit on your various committees in your career, now you want to forbid the use for culture war intra-elite signaling reasons.
You don't own the knowledge you put out there unless you have a limited time valid patent. The rest is absurdity. If you want to keep your findings to yourself, keep them secret.
The intellectual property law that governs ACM articles is copyright law, not patents. I don’t know who controls these (the ACM or the authors) or what rights might have been granted to the public.
The entire point of copyright law is so that people can make their writing public and still be able to control the right to make copies (for example, into your dataset for training an LLM).
Are you a lawyer that has tested this in court, or is this just what you want the reality to be?
As someone with lots of open source code out there that has likely been used as LLM training data, I'm very sympathetic to this point of view, but that doesn't seem to be the legal reality. Much of this has not been fully tested in court, but it seems likely that LLM training is not copyright infringement, as long as the training material itself was acquired legally.
I mean, its theoretically possible that a court might rule that if an LLM outputs an exact or lightly modified piece of copyrighted work, that it won't be copyright encumbered. We'll end up in a situation where copyright doesn't exist anymore, because you can always claim that its been laundered through an AI. This seems terribly unlikely to me, because 1:1 transformations (eg copying an image into memory) are already established to count as making a copy for legal purposes, there's strong precedent around piracy
There's also been court cases where material has been found to be infringingly used, eg song lyrics, so the case where copyright ceases to exist doesn't seem to be coming through yet, thankfully. It'd be the most staggering upheaval of copyright of all time if this doesn't turn out to be true
The verbatim reproduction is clearly a red herring and not the main use case. Nobody reads novels (or science papers) by prompting ChatGPT to give the next paragraph.
Derivative work or transformative? It's not the same.
AI code generation often outputs exact copies of code that exists in the wild. I've seen it output chunks from research papers unprompted as well, its a big problem, or blending two papers together in a salad
AI works also clearly aren't transformative in many cases. If you ask it a question about a paper, it'll quote bits of the paper at you. That serves as an exact substitute of the original work. If you ask it for song lyrics, or information about the news, its content is a direct substitute for the original source it was trained on. This clearly does not fall under a transformative use case
You could argue that some uses of it are transformative, but even then - its easy to find some piece of training data in the source code that the output work supersedes. By its very nature it does not have the capacity to genuinely invent under the law (as it is not human), and a prompt isn't a significant enough part of the processing to count here
As long as it is not substantially similar to training set it should be OK to reuse ideas. Ideas should not be protected by copyright or we can't create anything. We should not accept "vibe copyrights".
Everyone treats the human user of the AI as furniture, but they steer the whole process into unique directions.
Under the law, machines are not able to create copyright, and basic human involvement is not enough to change this. Eg if you click a button saying "go", that won't create copyrightable content
You can use all the ideas you want, but AI cannot because its not a person, and does not enjoy the same protection under the law. The copyright holders by and large did not agree to you using their content like this
If we enable this, people won't create anything because all their work will immediately be stolen by the AI models. Copyright partially exists to promote the creation of new content, because theft disincentivises novel creation
Are you sure you didn’t have RAG enabled and it wasn’t putting the paper in its context for your prompt? I find it hard to believe that anything but the most commonly published papers/code would exist directly in model weights.
1. AI models frequently output large chunks of code which are plagiarised. In one specific case it was code for walking the stack, that was a clear mix of two original sources that I was able to find with changed variable names, but the structure was identical and switched from the first to the second halfway through
2. AI models plagiarising stack overflow answers word for word, quite recently about the rotation rate of smoothbore cannons in the age of sail
3. AI misspelling answers because the physics papers its trained on made the same typos, which is how I discovered that it had plagiarised the answer
4. Misconceptions/wrong answers that can be traced back to specific papers due to the oddly specific nature of the language used
There's been a lot of research about getting AI models to output their training data, and it turns out they store huge amounts of it. You can use this to get people's personal information if you really want to, and that's very low occurance information
I just want to make sure that you saw those things with pure context isolation and AI wasn’t just doing a search to grab the content directly and throw it into the context of your prompt. You’d be surprised how often I’ve heard these claims and it turns out they were just using an agent with RAG enabled.
AI has the capacity to exactly reproduce its training data, just because something is a transformed representation does not mean that it isn't copying it in some fashion. The JPEG format 'just' counts the frequencies in an 8x8 block of pixels, and yes that's 100% copyright infringement
I have the capacity to exactly reproduce things I've read as well, but it's not automatically copyright infringement if I do so.
> and yes that's 100% copyright infringement
Says what court of law?
I'm kinda getting tired of this stuff. I'm someone who has been, and still to some extent is, uncomfortable with the possibility of copyright/license laundering in LLMs, but they way you are making your argument is incredibly off-putting and not sympathetic. You're throwing out wild assertions about the law that are not supported by... anything, really.
There's way too much hand waving on this topic. It is legal to produce copywritten work. If I draw Pikachu the drawing is mine. Legally. I am simply unable to make money on it. I can give it away if I want with zero liability. I could even hang the drawing up in my restaurant as a decoration. No big deal. What I can't do is use that drawing as my mascot or branding. We have an entirely separate process to determine if you are infringing on a copyright / trademark by using it to sell something. That's why whether or not an LLM can produce a picture of Pikachu is largely irrelevant. It's what you do with it that matters. Even more interestingly if I draw a picture of Pikachu and then the Pokemon Company decides they wanna use that specific picture they actually would have to pay ME for the copyright to use it.
Let's not mix copyright and trademark in mixed phrases like "infringing on a copyright / trademark". The two are very different concepts with different goals.
The main question in the "AI image generator generates a Pikachu image" is whether the AI company serving that image generator to you is violating the copyright or not. Because they make money when doing so (API / subscription cost), and so it's like selling images of Pikachu. The user is likely in the clear as long as they don't go on sell that Pikachu further. But the AI company sold the Pikachu image to the user.
That question is irrelevant. Artists may be hired to reproduce copywritten work without the consent of the copyright owner. In this case an LLM is no different from Photoshop. It is a tool. Nothing more.
Fair use is a defence against copyright infringement. Ie you actively say that you *have* committed copyright infringement, but you're allowed to do it under fair use doctrine to train the model. That says nothing about the purposes the model is used for
There's also these parts:
> its use of pirated books to create such library does not constitute fair use.
Which indicates that there are tight bounds depending on the ethics of how the content was obtained
Similarly with the second one
>Meta moved to dismiss plaintiffs’ cause of action for direct copyright infringement only to the extent that it was premised on a theory that the software comprising LLaMA is itself an infringing derivative work.
We're talking specifically about the output of the models being infringing, not whether or not the models themselves are infringing. If you read onwards
>Plaintiffs’ claim for vicarious copyright infringement failed because the complaint did not allege that any output generated by LLaMA contained protectable expression that recast, transformed or adapted the books. Without “an infringing output, there can be no vicarious infringement.”
Which strongly indicates the precise opposite of what you're saying, if you actually like, read the rulings
I don't think exact copies are what anybody is worried about. It really is the information itself. The GP's been lulled into thinking he owns the knowledge when he only owns (some) rights to the creative way he wrote it. It really is objectionable of him to want to restrict access. Intra-elite culture war is the perfect term to describe it.
The world would be quite different if AI companies had to create the knowledge they trained on, rather than consume that knowledge freely given away. They're not known for freely giving away their produce either, I don't know why you think the ire should be pointing in this direction.
I like open models for sure. I support free software as well. But using published knowledge to solve new problems was never disallowed, even for profit. Today, a for-profit company, e.g. a gigantic Big Pharma company can have their employees read chemistry and biology papers and use the knowledge gained from it to improve their products and processes and make more profit without paying a cent to the authors (beyond what they may get - likely nothing - due to the potential paywall).
This makes no sense whatsoever. How would a philosopher of science, or a social scientist who publish in the ACM apply for a patent? Or someone who builds software (software patent not so easy to get ;)). I have applied for patents before and I'm pretty sure my patent application has been fed to countless LLMs by now.
The issue is not who owns knowledge, it's how it benefits humanity.
They can't apply for a patent in those categories, so they have no way of preventing others from reading their text (or using a program to process the text, and compute statistical properties), and using the ideas in other contexts (without re-expressing the same text).
If someone reads a philosophical essay, has a heureka moment from it, applies the principle to their work, and makes bank (commercial profit), they never have to pay a percentage to the author of the essay.
Patents were invented to incentivize disclosure of inventions and prevent extended secrecy. In exchange for the public disclosure (which allows others to experiment with the idea without selling yet), you get to keep exclusivity for N years.
“If you don’t lock your bike, you can’t prevent others from taking it for a ride. You enjoy the mobility attached to the idea you’re a bike rider who rides a bike but then you play this game when that bike would actually be useful as opposed to sitting in the bike rack all day”.
Copycats copy it, then it gets associated with subpar quality then the frontier guys adopt something fresh (or something deliberately currently-very-janky or retro/outdated associated) and the barber's pole of fashion rotates once more.
Poking one hole in the walls will always be easier than guarding the entire frontier.
To defend properly, you need either either much more formally mathematically verified layers, or always-on online monitoring and intrusion detection running alongside the service. Like the immune system, or like guarding an empire's borders, since you can't put a guard on every meter of border. But you can detect when someone broke in and deploy your soldiers to respond. Currently we are trying to defend by building really hard walls that we hope cannot be pierced anywhere, which isn't realistic.
This is of course about the PhD level and beyond. For regular university-level education, where the goal is to pass the usual exams, ChatGPT and Claude are more than enough. But for real-world "doing", not (yet).
reply