Google has no contract with OpenAI though. They used a third party site to scrape conversations. If the outputs themselves are not copyrighted, and they never agreed to the terms of service, it should be fine, right? Albeit unethical and embarrassing.

> Albeit unethical and embarrassing.

I really don’t understand this angle. In fact, I am fairly positive that the training set for GPT-4 contains many thousands of conversations with AI agents not developed by OpenAI.

Do AI companies need to manually sift through the corpus and scrub webpages that contain competitor LLM output?

(“Yes” is an acceptable answer to this, but then it applies to OpenAI’s currently existing models just as much as to Bard)

How did you come about being "fairly positive" that GPT-4 is trained on other AI conversations?

Many AI conversations have been floating around internet forums since the original GPT was released. As OpenAI hasn't shared anything about its training set, to err on the side of caution I would assume that they didn't filter these conversations out. If they aren't even marked as such, it may not even be possible to do. I think it would be very hard to prove that no AI conversations are included in the training set, even if it wasn't secret.

Hardly unethical, considering OpenAI is doing exactly this.

You can quibble about the ethics of web scraping for ML in general but I think you're conflating issues.

OpenAI and Google both scour the web for human-generated content. What Google cares about here is the learnings from OpenAI's proprietary RLHF dataset, for which they had to contract a large sum of human labelers. Finding a roundabout way to extract the value of a direct competitor's purpose-built, costly data feels meaningfully different from scraping the web in general as an input to a transformative use.

> OpenAI and Google both scour the web for human-generated content

OpenAI and Google both scour the web for content, period. That content could be human generated or AI generated or a mix of the two. Neither company is respecting copyright or terms of service of every individual bit of data collected. Neither company cares how much effort was put into creating the data, whether humans were paid to do it, or whatever else. So there really isn't that much difference between the two.

And herein is the main problem of AI. Its creators consume knowledge from the commons, and give nothing free and unencumbered back.

It's like the guy who never brings anything to the potluck, but after everyone finishes eating, he boxes up the leftovers, and starts selling them out of a food cart.

> after everyone finishes eating, he boxes up the leftovers, and starts selling them out of a food cart.

That particular example doesn't seem all that great, as it gives the impression of not letting food go to waste.

Though sure, the guy could just give the food away, but shrug.

So what? Is OpenAI RLHF dataset more valuable than millions of books and paintings OpenAI used for free without stopping a second? Why is that? Because one big tech corp paid money for that dataset?

If there's a party which has intentionally conflated scraping web content in general with scraping it to build a direct competitor to the original sources, that party is Google.

Yes, this latest instance with OpenAI outputs is shady, but I think it's in the same spirit as scraping news organizations for content which journalists were paid to write, and then showing portions of it directly in response to queries so people don't go directly to the news organization's pages, and it's in the same spirit as showing answers to query-questions that are excerpts from scraped pages which another organization paid to produce.

I see no difference. Any web scraping is a means to deflect revenue-generating traffic to yourself, and away from other websites. Fewer people will go to Stack Overflow because of Codex and Copilot. The point that the content was paid for vs volunteered becomes moot once it's posted publicly online for free, on ShareGPT.

> labelers. Finding a roundabout way to extract the value of a direct competitor's purpose-built, costly data feels meaningfully different from scraping the web in general as an input to a transformative use

There we go again, its, one law for the unwashed plebs and the other for us.

Why do you think that I, after spending my time and effort to write my blog, own my content to a lesser extent that OpenAI does their? Such hypocracy.

Two wrongs don’t make a right.

forgive me if i have limited sympathy when a burglars house gets robbed

This

It's even less worthy of sympathy - like a counterfeit piece of art being counterfeited. And there isn't even an original, just like a made up counterfeit.

That's not how rights work, though. I'm sure you don't want your rights to be conditional on the sympathy of others.

good thing I'm just some guy expressing an opinion, and not a judge, then

It’s still debatable if training a computer neutral network on public data is 'wrong' when we very much accept it as a right for biological neural networks.

No more unethical or embarrassing than scraping the web for millions of copyrighted works and selling access to unauthorized derivative works.

The recent HiQ vs LinkedIn case would seem to make this ToS unenforceable, unless Google actually created a user account on ShareGPT and affirmatively accepted the terms. "Acceptance by default" does not count, and I can easily browse ShareGPT without affirmatively accepting any ToS, without which web scraping is totally legal.

So, to verify, are you claiming that if someone added a similar clause to their source code and then GitHub went ahead and trained Copilot against it, that would be an issue?

You relinquish all licensing rights when you upload your code to GitHub. Microsoft can do whatever they want with it. That's in their ToS, which you have to agree to when you make an account. Normally, only affirmatively accepted ToS are enforceable, so just putting a clause into your license doesn't work (unless it's a copyright, which doesn't require consent).

> You relinquish all licensing rights when you upload your code to GitHub

What now? Seriously?

I found this. Section D4.

"We need the legal right to do things like host Your Content, publish it, and share it. 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 includes the right to do things like copy it to our database and make backups; show it to you and other users; parse it into a search index or otherwise analyze it on our servers; share it with other users; and perform it, in case Your Content is something like music or video."

"as necessary to provide the Service" seems critical.

Also, section D3 of the GitHub Terms of Service says:

> You retain ownership of and responsibility for Your Content.

and section D4 says:

> 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.

There is nothing in the terms that requires the GitHub user to relinquish all licensing rights.

https://docs.github.com/en/site-policy/github-terms/github-t...

The clauses always have a trap door: "[outside of] our provision of the Service" means they can do anything as long as it's a service they provide.

Under definitions: The “Service” refers to the applications, software, products, and services provided by GitHub, including any Beta Previews.

I think there's misunderstanding over what the word "relinquish" means.

The terms make clear that uploading code to GitHub gives GitHub 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" while the code is hosted on GitHub.

However, that's not the same thing as relinquishing (giving up) licensing rights to GitHub. The uploader still retains those rights, and there is nothing in the terms that says otherwise.

The question turns on whether you consider copilot part of the "GitHub service."

GitHub would argue that it is, and they'd likely argue that charging for access to copilot is akin to charging for access to private repositories.

Others would say that copilot is somehow separate from the services Github provides, so using their code for CoPilot wouldn't be covered by the ToS.

It is certainly a service that's being provided. If not by GitHub, then by whom?

I'll repeat the definition of service: The “Service” refers to the applications, software, products, and services provided by GitHub, including any Beta Previews.

So do you believe if you hosted a closed source project on GitHub, and GitHub decided they want to integrate this into their service they would simply be allowed to take the code?

Fortunately HN commenters are not judges. And I would wager any bet that MS lawyers would not try to argue based on their ToS either, that would be a recipe for loosing any court case.

I just mean that it doesn't really matter what your license says as long as GitHub can come up with a business justification for using it in some way. Certainly, other users still legally have to obey your copyright.

"Improving the service over time" can do a lot of heavy lifting, definitely including training Copilot.

So, to verify, are you claiming it would not be allowed for you to upload my otherwise-open-source code (code I do not myself host at GitHub, but which was reasonably popular / important code) to GitHub?

Yep. It's in their ToS:

If you're posting anything you did not create yourself or do not own the rights to, you agree that you are responsible for any Content you post; that you will only submit Content that you have the right to post; and that you will fully comply with any third party licenses relating to Content you post.

I suppose this means if I upload your stuff to GitHub, and you sue GitHub, then GitHub would be able to somehow deflect liability onto me.

That doesn't make sense. For example, GPLv3 allows anyone to redistribute the software's source code if the license is intact:

> You may convey verbatim copies of the Program's source code as you receive it, in any medium, provided that you conspicuously and appropriately publish on each copy an appropriate copyright notice; keep intact all notices stating that this License and any non-permissive terms added in accord with section 7 apply to the code; keep intact all notices of the absence of any warranty; and give all recipients a copy of this License along with the Program.

https://www.gnu.org/licenses/gpl-3.0.en.html

If GitHub uses the source code in a way that violates the license, there is no provision in the GitHub terms of service that would allow GitHub to deflect legal liability to the GitHub user who uploaded the program. The uploader satisifed the requirements of GPLv3, and GitHub would be the only party in violation.

Uploading is granting GitHub a license separate from the gpl license.

If you can't actually grant that separate license, you're misrepresenting your ownership and license to that code

I'd like to see that theory tested in court. Section D3 of the terms says:

> If you upload Content that already comes with a license granting GitHub the permissions we need to run our Service, no additional license is required.

https://docs.github.com/en/site-policy/github-terms/github-t...

and section D4 does not mention any permissions that GPLv3 does not already cover. GitHub automatically recognizes when a repo is GPLv3-licensed, so it cannot claim ignorance of what GPLv3 is.

I love it how they don't want others to use their model output but they have no qualms about training their model on the copyrighted works of others? Isn't this a stunning level of hypocrisy?

breaking terms of service is not punishable in any way. Facebook tried and lost in court

Correction – breaking terms of service that you have not explicitly agreed to is not punishable in any way. A site cannot enforce a "by using this site you agree to..." clause deep inside some license page that visitors are generally unaware of. If you violate an agreement that you willingly chose to enter, however, you will likely be found liable for it.

Where exactly does it do that? I looked a bit and could t find it, but likely I was just wrong

Are the TOS even enforceable is AI content can't be copyrighted?

This is really hilarious. Authors and artists never gave permission to use their work to train AI models either...

Not legally the same situation, but ethically close enough.