3 ms·

So? First off, the whole argument behind these models has been from day one that training on copyrighted material is fair use. At most this would be a TOS viol

So? First off, the whole argument behind these models has been from day one that training on copyrighted material is fair use. At most this would be a TOS violation. Second off, AI output is not subject to copyright, so it has even less protection than the original works it was trained on. Copyright maximalism for me, but not for thee. It's just so silly for someone working at OpenAI to complain about this.

robocat 4y ago
> AI output is not subject to copyright The chats include human output too, which is presumably copyrighted, and is presumably necessary for training purposes.
- danShumway 4y ago
  OpenAI doesn't own the copyright on the human aspects of the chats, so it still doesn't really have a claim to make around them. And even if it did own that copyright, we loop right back around to "wait, training an AI on copyrighted material isn't fair use now?" There's no way that ChatGPT's conversations are going to be subject to more intellectual property protection than the human chats it was trained on.
magicalist 4y ago
> At most this would be a TOS violation And would it be a ShareGPT TOS violation (assuming it had any)? If OpenAI says "you can share these online but don't use them for AI training", people share them on another site, and then someone else comes along to scrape that site for AI training data, there's no relationship between OpenAI and the scraper for the TOS to apply to. Normally I think you'd rely on copyright in that kind of case, but that doesn't apply to ChatGPT's output, so...
- danShumway 4y ago
  Right. And what even is the penalty of that TOS violation and how enforceable is it? I don't have an OpenAI account. I have never agreed to any TOS. I don't see what legal claim they would have to stop me from training an LLM on ShareGPT.
  - judge2020 4y ago
    If Google were specifically going to ChatGPT to get its output and train off of it, they could be sued for breach of contract - and OpenAI would likely have a pretty good argument: - they specifically tried extracting and learning from our model when it says you can't in our TOS - this makes it easier for them to compete with us via the data they obtain in their breach of contract - more businesses and enterprises might pass up on renting a shared or dedicated instance from us if they can just get it from Google
    - danShumway 4y ago
      > If Google were specifically going to ChatGPT to get its output and train off of it But (correct me if I'm wrong) I don't think anyone anywhere is claiming that's what happened. The claim was just that Google looked into using existing chats that it scraped from another website. Edit: realizing you're probably replying specifically to the question I asked, "and what even is the penalty of that TOS violation and how enforceable is it?" In which case, yeah, that's a decent clarification to add, sorry for pushing back on it.
yreg 4y ago
> It's just so silly for someone working at OpenAI to complain about this. Who from OpenAI is complaining?
- danShumway 4y ago
  My understanding is that the Twitter thread author works at OpenAI. Maybe I'm wrong about that.
  - yreg 4y ago
    According to his bio, he works at Vercel. He made a hobby project called ShareGPT[0] and that's probably where the accusation came from. [0] - https://sharegpt.com/ https://sharegpt.com/
    - danShumway 4y ago
      Thanks for the clarification. Aside from the OP, I haven't seen anyone from OpenAI commenting on this, so yeah, unless I've missed something I think you're correct to point that they're not involved so far.

robocat 4y ago

> AI output is not subject to copyright The chats include human output too, which is presumably copyrighted, and is presumably necessary for training purposes.

danShumway 4y ago
OpenAI doesn't own the copyright on the human aspects of the chats, so it still doesn't really have a claim to make around them. And even if it did own that copyright, we loop right back around to "wait, training an AI on copyrighted material isn't fair use now?" There's no way that ChatGPT's conversations are going to be subject to more intellectual property protection than the human chats it was trained on.

magicalist 4y ago

> At most this would be a TOS violation And would it be a ShareGPT TOS violation (assuming it had any)? If OpenAI says "you can share these online but don't use them for AI training", people share them on another site, and then someone else comes along to scrape that site for AI training data, there's no relationship between OpenAI and the scraper for the TOS to apply to. Normally I think you'd rely on copyright in that kind of case, but that doesn't apply to ChatGPT's output, so...

danShumway 4y ago
Right. And what even is the penalty of that TOS violation and how enforceable is it? I don't have an OpenAI account. I have never agreed to any TOS. I don't see what legal claim they would have to stop me from training an LLM on ShareGPT.
- judge2020 4y ago
  If Google were specifically going to ChatGPT to get its output and train off of it, they could be sued for breach of contract - and OpenAI would likely have a pretty good argument: - they specifically tried extracting and learning from our model when it says you can't in our TOS - this makes it easier for them to compete with us via the data they obtain in their breach of contract - more businesses and enterprises might pass up on renting a shared or dedicated instance from us if they can just get it from Google
  - danShumway 4y ago
    > If Google were specifically going to ChatGPT to get its output and train off of it But (correct me if I'm wrong) I don't think anyone anywhere is claiming that's what happened. The claim was just that Google looked into using existing chats that it scraped from another website. Edit: realizing you're probably replying specifically to the question I asked, "and what even is the penalty of that TOS violation and how enforceable is it?" In which case, yeah, that's a decent clarification to add, sorry for pushing back on it.

yreg 4y ago

> It's just so silly for someone working at OpenAI to complain about this. Who from OpenAI is complaining?

danShumway 4y ago
My understanding is that the Twitter thread author works at OpenAI. Maybe I'm wrong about that.
- yreg 4y ago
  According to his bio, he works at Vercel. He made a hobby project called ShareGPT[0] and that's probably where the accusation came from. [0] - https://sharegpt.com/ https://sharegpt.com/
  - danShumway 4y ago
    Thanks for the clarification. Aside from the OP, I haven't seen anyone from OpenAI commenting on this, so yeah, unless I've missed something I think you're correct to point that they're not involved so far.