12

This is a subset of What, exactly, is Stack Overflow's agreement with OpenAI?, but it seems that we are playing Twenty Questions this month.

  • Representatives of the company (e.g. the Chief Product Officer, the Director of Product Innovation, Rosie) have repeatedly stated that socially-responsible purveyors of AI models must provide attribution. Most recently,

    Having credit attributed is a non-negotiable for us, and is a critical part of any and all partnerships of this type.

  • The company has not tried to re-license our contributions:

    In our agreements with our partners, we have stated that the data they use is governed by the Creative Commons license.

  • The company does not require that its OverflowAI partners (currently Google and OpenAI) abide by these requirements in any particular way:

    We are not solutioning (aka, directing or telling) how our partners meet these requirements for attribution and appropriate credit.

With that said:

  • Does the company require that OverflowAI partners:
    • only use Stack Exchange contributions in accordance with the relevant CC BY-SA license, and
    • not use Stack Exchange contributions in ways that contravene the license?
  • If it does require this, does the company have tools to enforce this requirement?
    • If so, what are they?
    • If not, why not?
  • If it does not require this, will this oversight be corrected?
    • If so, how and when?
    • If not, why not?
Improve this question
10
  • 5
    another question that they will not answer I fear. But at least, the question will be here for everyone to see. Commented May 17, 2024 at 17:53
  • 3
    I posted this on basically the weekend, so please add 3 days to your "hasn't responded" threshold before getting upset. Management making poor decisions shouldn't create an expectation for blameless staff to work overtime. Commented May 17, 2024 at 18:12
  • 2
    Are you asking about terms similar to the AUP, which allows the company to block access from scrapers who are violating the license? Because the company cannot directly enforce the license itself. Commented May 17, 2024 at 18:25
  • 2
    @ThomasOwens I'm asking about the deals that Stack Exchange has made with its partners. The company cannot enforce the license in the absence of another agreement, but there's no reason a contract couldn't include "you have to do XYZ" for near-arbitrary XYZ. Commented May 17, 2024 at 22:06
  • 3
    It's unfortunate, but I don't think this is asking the right questions; I think Q1 is answered wholly by Rosie's post (i.e. "yes"), and Q2 is "no" for the same reason that SE stopped taking reports of scraping requests: they can't enforce your copyright. What I'd like to see instead is some sort of material commitment to action in the case that some partner does provably break the license of our content, e.g. terminate their agreement with that partner. It'd still just be words, but it'd be actionable and specific, which might help assuage some of the community's concerns. Commented May 18, 2024 at 0:32
  • @wizzwizz4 forgive me to be pedantic, but since you had to point that out first... Saying "they will not answer I fear" implies a future event, not a past one. You can fear something happening, but once it happened it is no longer just a fear, it becomes a fact. That message was not about an already past event - it was about a behavior pattern I have noticed. So I guess I have nothing to "add", since I didn't say "they didn't answer (present/past)", I said that I expected (fear) "they will not answer (future)". I'll be glad to be wrong, see you in 3 days then. Commented May 20, 2024 at 7:46
  • 2
    @berthold have you considered renaming the tag instead of retagging the individual questions? Commented Sep 30, 2024 at 15:56
  • 1
    @Catija it was moving some posts from one tag to another, but not all. Retagging the ones that needed a different tag seemed like the best option. Commented Sep 30, 2024 at 16:08
  • 2
    Do we have any examples, in practice, of the AI companies that have been whitelisted to crawl SO and use it's content properly citing that content when in use? Commented Sep 30, 2024 at 16:22
  • 1
    @KevinB I'm not aware of any. Might be worth asking that as a different question. (This is already a subset of a different question.) Commented Sep 30, 2024 at 16:30

0

Reset to default

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.