Skip to main content

You are not logged in. Your edit will be placed in a queue until it is peer reviewed.

We welcome edits that make the post easier to understand and more valuable for readers. Because community members review edits, please try to make the post substantially better than how you found it, for example, by fixing grammar or adding additional resources and hyperlinks.

7
  • 1
    I wonder if it's possible to legally block AI training from using CC BY SA licensed content (which includes the entirety of SE). As I understand it, if they attribute the source, they can use the content. If you disagree, the only way to find out who's right is probably to sue. (Do they attribute their sources? Not that I've found, but if that's the only legal qualm then they could continue what they're doing if they properly attributed their sources.)
    – Laurel
    Commented Dec 13, 2022 at 16:13
  • 1
    @Laurel question is how they scrape the training data, in what tools, etc. Commented Dec 13, 2022 at 16:19
  • @Laurel law.stackexchange.com/q/11183/31 Commented Dec 16, 2022 at 17:20
  • @FranckDernoncourt's link (somewhat) addresses the attribution side of things. However, that's not the core of the issue. Stack Exchange posts are licensed under CC BY SA, which cannot be revoked. You can tell someone they can't use your stuff, but it's an empty threat legally if you've licensed it under a license that allows that use. Or was this suggesting a new license for future contributions? Are there any licenses out there like that or is it unenforceable?
    – Laurel
    Commented Dec 16, 2022 at 17:56
  • @Laurel "In the case of a share-alike-licensed dataset, must all models trained on it be redistributed under the same or similar license?" Commented Dec 16, 2022 at 17:57
  • @FranckDernoncourt And if they distribute it under a compatible license?
    – Laurel
    Commented Dec 16, 2022 at 18:02
  • @Laurel I think it's ok, unless perhaps if the license excludes derivative work. Commented Dec 16, 2022 at 18:04