It sounds fair from how the article describes it
That said the implementation doesn't appear to be totally trivial and copilot apparently even copies the comments which are almost certainly copyrightable in themselves.
https://x.com/StefanKarpinski/status/1410971061181681674 https://github.com/id-Software/Quake-III-Arena/blob/dbe4ddb1...
However a twitter post on its own isn't evidence a court will accept. You would need the original poster to testify that what is seen in the post is actually what he got from copilot and not just a meme or joke that he made.
Also the plaintiffs in this case don't include id-Software and there is some evidence that id-Software actually stole the fast inverse sqrt code from 3dfx so they might not want to bring a claim here anyways.
When it was reported, I was able to reproduce it myself.
Absolutely there were a few outliers where a judge might want to look more closely. I'd be surprised if -under scrutiny- there wouldn't be any issues whatsoever that OpenAI overlooked.
However, it seemed to me that over half of the NYT complaints were examples of using the -then rather new- ChatGPT web browsing feature to browse their own website. In the case, they then claimed surprise when it did just what you'd expect a web browsing feature to do.
i agree from a philosophical pov, but this is clearly not the case in law.
https://en.wikipedia.org/wiki/Abstraction-Filtration-Compari...
All the plaintiffs would need to do is provide evidence that copywritten code was produced verbatim. This includes showing the copyrighted code on GitHub, showing copilot reproducing the code (including how you manipulated copilot to do it), showing that they match, and showing that the setting to turn off reproduction of public code is set.
It makes no difference who owns the copyrighted code, it need only be shown that copilot is violating copyright. Microsoft can't say "uhh that doesn't count" or whatever simply because they own a company that owns a company that owns copyright on the code.
Rightly so, you have to show some sort of damage to sue someone, not just theoretical damages.
Open source licenses allow sharing under certain conditions.
Also, even if this weren’t the case you can’t sue for damages to other people (they’d need to bring their own suit)
Deleted comment
It would be more correct to say Quake III Arena was released to the public as free software under the GPLv2 license.
Copyright infringement could be emitting the code in a manner that exceeds fair use.
The license gives you permission to utilize the code in a certain way. If Copilot gives you GPLed code that you then put into your closed source project, you have infringed the license, not Copilot.
> If you don't meet the conditions, it's still copyright infringement like before.
Licensing and copyright are two separate things. Neither has anything to do with the other. You can be in compliance with copyright, but out of license compliance, you can be the reverse. But nothing about copyright infringement here is tied to licensing.
To be clear: I am a person who trashed his Reddit account when they said they were going to license that text for training (trashed in the sense of "ran a script that scrubbed each of my comments first with nonsense edits, then deleted them"). I am a photographer who has significant concerns with training other models on people's creative output. I have similar concerns about Copilot.
But confusing licensing and copyright here only muddies waters.
It'd be a long bow to draw to say that what is akin to a search result of a snippet of code is "redistributing a software package".
1. The copilot team rushed to slap a copyright filter on top to keep these verbatim examples from showing up, and now claims they never happen.
2. LLMs are prone to paraphrasing. Just because you filter out verbatim copies doesn't mean there isn't still copyright infringement/plagiarism/whatever you want to call it. The copyright filter is only a legal protection, not a practical protection against the issue of copyright infringement.
Everyone who knows how these systems work understand this. The copilot FAQ to this day claims that you should run copyright scanning tools on your codebase because your developers might "copy code from an online source or library".
Github has it's own research from 2021 showing that these tools do indeed copy their training data occasionally: https://github.blog/2021-06-30-github-copilot-research-recit...
They clearly know the problem is real. Their own research agreed, their FAQs and legal documents are carefully phrased to avoid admitting it. But rather than owning up to the problem, it's "Ner ner ner ner ner, you can't prove it to a boomer judge".
More than that: the fact that they claimed it wasn't possible before adding the filter, to filter out the thing that said wasn't possible. This doesn't help me trust anything else they might say or have already said.
My take on that was always: if it isn't possible, then why are MS not training the AIs on their internal code (like that for Office, in the case of MS with their copilot product) as well as public code? There must be good examples for it to learn from in there, unless of course they thing public code is massively better than their internal works.
Since you really need to work hard to make the AI spit out anything verbatim, and you have no knowledge of their internal code, how could you ever prove or deny it?
Because if they were, they would have said.
It would be an excellent answer to the concerns being discussed here: “we are so sure that there is nothing to worry about in this regard, that we are using our own code as well as the stuff we've schlepped from github and other public sources”.
Actually, it does. The production of the output is what matters here.
People do clean room implementations because of paranoia, not because it's actually a necessary requirement.
The literal act of making modifications isn't infringement until you distribute those modifications -- and we're talking about a situation where you've changed the code enough that it isn't considered a derivative work anymore (apparently) so that's kosher.
> you already have license to access the code
This isn’t access, that occurs before the AI is trained. It’s access > make copy for training > AI does lossy compression > request unzips that compression making a new copy > process fuzzes the copy so it’s not so obvious > derivative work sent to users.
>You grant us and our legal successors the right to store, archive, parse, and display Your Content, and make incidental copies, as necessary to provide the Service, including improving the Service over time
>This license does not grant GitHub the right to sell Your Content. It also does not grant GitHub the right to otherwise distribute or use Your Content outside of our provision of the Service, except that as part of the right to archive Your Content, GitHub may permit our partners to store and archive Your Content in public repositories in connection with the GitHub Arctic Code Vault and GitHub Archive Program.
https://docs.github.com/en/site-policy/github-terms/github-t...
I think the important questions are (1) whether "the Service" includes Copilot, and (2) whether GitHub is selling users' content with Copilot.
For (1), I'm unhappy to admit Copilot probably does fall under "the Service," which is nebulously defined as "applications, software, products, and services provided by GitHub." But I'll still say that users' could not agree to this use while GitHub was training The Copilot model but hadn't yet announced it. At that time, a reasonable user would've believed GitHub's services only covered repository hosting, user accounts, and the extra features attached to those (issue trackers, organizations, etc).
GitHub could defend themselves on point (2) by saying they aren't selling the code, instead selling a product that used the code as input. But does that differ much from selling an online service that relies on running user code? The code is input for their servers, and it doesn't need to be distributed as part of that questionable service. But it's a clear break from the TOS.
GitHub didn’t just copy open source code they copped everything without respect to license. As such attribution which may have allowed some copying isn’t generally relevant.
Really a public repo on GitHub doesn’t even mean the person uploading it owns the code, if they needed to verify ownership before training they couldn’t have started. Thus by necessity they must take the stance that copyright is irrelevant.
If you copy a whole book and do the same, there’s still lines-3 infringement left.
Isn't that akin to destruction of evidence?
In spirit? ... Probably?
Unlike most LLMs, Github copilot can trivially solve their copyright problem by just using only code they have the right to reproduce.
They have a giant corpus of code tagged with license, SELECT BY license MIT/Equivalent and you're done, problem solved because those licenses explicitly grant permission for this kind of reuse.
(It's still not very cash money to take open source work for commercial gain without paying the original authors, and there's a humorous question if MIT-copilot would need to come with a multi-gigabyte attribution file, but everyone widely agrees it's legal and permitted.)
The only reason you'd hack a filter on top rather than doing the above is if you'd want to hide the copyright problem. It's an objectively worse solution.
Absolutely not trivial, in fact completely impossible by computer alone. You can't determine if you have the right to reproduce a piece of code just by looking at the code and tags themselves. *Taps the color-of-your-bits sign.*
* I can fork a GPL project on Github and replace the license file with MIT. Okay to reproduce?
* If I license my project as MIT but it includes code I copied inappropriately and don't have the right to reproduce myself, can Github? (No) This one is why indemnity clauses exist on contracted works.
* I create a git repo for work and select the MIT license but I don't actually own the copyright on that code and so that license is worthless.
Is taking away a drunk driver's keys (before they get in the car) destruction of the evidence of their drunk driving?
In the current case - its unclear if any crime took place at all, it seems clear that the primary intent was to prevent future crime not hide evidence of past ones. Most importantly the past version of the app is not destroyed (presumably). Github still has the version of the software without the copyright filter. If relavent and appropriate, the court could order them to produce the original version. It can't be destroying evidence if the evidence was not destroyed.
The people that think Copilot is infringng their copyright would be happy with that I would think? Unless they take a much stricter definition of fair use than current courts do.
Well if the copyright filter is working they indeed aren't happening. Putting in safe gaurds to prevent something from happening doesn't mean you're guilty of it. Putting a railing on a balcony doesn't imply the balcony with railing is unsafe.
> LLMs are prone to paraphrasing. Just because you filter out verbatim copies doesn't mean there isn't still copyright infringement/plagiarism/whatever you want to call it
Copyright infringement and plagerism are different things. Stuff can be copyright infringement without being plagerized, and can be plagerized without being copyright infringement. The two concepts are similar but should not be conflated, especially in a legal context.
Courts decide based on laws, not on gut feeling about what is "fair".
> They clearly know the problem is real
They know the risk is real. That is not the same thing as saying that they actually comitted copyright infringement.
A risk of something happening is not the same as actually doing the thing.
> "Ner ner ner ner ner, you can't prove it to a boomer judge".
Its always a cop-out to assume that they lost the argument because the judge didn't understand. I suspect the judge understood just fine but the law and the evidence simply wasn't on their side.
Doesn't mean you weren't, at some point, guilty of it, either. It doesn't retcon things.
After all, you yourself probably cannot prove that you didn't commit the same offense at some point in time in the past. Like Russel's teapot, its almost always impossible to disprove something like that.
The main issue, as I see it, is that they took copyrighted material and made new commercial products without compensating (let alone acquiring permission from) the rights holders, ie their suppliers. Specifically, they sneaked a fair use sticker on mass AI training, with neither precedent nor a ruling anywhere. Fair use originates in times before there were even computers. (Imo it’s as outrageous as applying a free-mushroom-picking-on-non-cultivated-land law to justify industrial scale farming on private land.) That’s what should be challenged.