26

Recently I was poking at arc browser. While my impressions of the browser are out of the scope of this question, their initial set up included an opt in set of AI features, and a rather thoughtful privacy page. I'll include a partial screenshot below should their page go down or change

enter image description here

I'm not enough of a privacy expert to say if this is a gold standard but as a potential end user - this seems very user centric, and clear

It includes what the feature is, what, when and why it's processed and by who and where data is processed. Considering the political climate we live in, there may be certain countries we wouldn't want our data going to, and having the information to make that determination would be nice. In terms of Teams, there might also be clients who may have restrictions on what/where their data can go.

Stack Overflow Inc. seems to be fairly determined to try to include generative AI in various products, both public platform wise, and on Teams. While I don't believe that it's the best path forward, it seems like a fair thing to ask on where our data is going and why.

I think of a few specific points to consider

  • documenting data flows for features the company intends to include in public Q&A. I'd note with some proposals, data may be sent before any forms of submissions

  • Teams has an optional AI feature. Considering this data is private, knowing what's going where might be reassuring to potential customers, especially considering potential regulatory issues

  • OverflowAPI might be tricky. At the very least, treating it as a subprocessor may be good. That said, considering SE is very vocal about these deals, I'd find it surprising if their customers are secret.

So, if the company is going to go forward on selling/using our data for/on AI, could we have some clarity on where it's going and why?

6
  • 5
    @SPArcheon Perhaps the more often these questions are asked, the less they are able to ignore them :)
    – Joachim
    Commented Oct 29 at 10:21
  • How do you define "our data" in this context? Some content, like questions and answers, have a CC BY-SA license, so that anyone can share and use it with conditions, is it still our data? About selling data, there was a question on this subject: meta.stackexchange.com/questions/401506/…
    – A.L
    Commented Oct 29 at 21:20
  • I'd say posts are a key part of my concerns here - both in terms of the 'post drafting experience' as well as SE's products that sell post data (aka overflowAI). And CC dosen't mean we've given up our copyright/moral rights to the posts. Commented Oct 29 at 22:54
  • FWIw, posts aren’t CC until posted. So for stuff like drafting assistants that’s still "your data". Commented Oct 30 at 4:44
  • 2
    Yup, and that's a slightly potentially funnly problematic place. What if I post PII or credentials then sanitise it for example. Commented Oct 30 at 6:46
  • 1
    @JourneymanGeek I had tried to kick start that argument too... But I am "antagonizing", so I kinda no longer care and will gladly let you inherit the One For All this time. Fell free to build up on or reference this Commented Oct 30 at 11:57

0

You must log in to answer this question.

Browse other questions tagged .