-8

SO/SE is a great resource for folks. These folks often [inadvertently] share information they might not have meant to.

A good example is when someone posts a question in any of the communities with a link to a Google document. Doing so exposes their email address. It's not easily obvious what the email address is, but it is exposed.

So the question is, when folks submit posts that unobviously expose such information, does SE/SO have a responsibility to alert the user?

Related discussion on the topic: Is it ethical to ask for a Google sheets file, when answering a question, even when such request violates a user's privacy?.

5
  • 4
    The SE network is a collection of sites to build a body of knowledge by users who are at least 13 years old. We assume they got their internet-101 elsewhere. It is not something SE (or any site for that matter) has to cater for.
    – rene
    Oct 6, 2022 at 20:18
  • 1
    API keys and passwords can be redacted (takes 2 moderators or a team member), but scrapers can be fast enough for it to have not enough effect. Some of the SE content is copied all over the place, even though SE takes steps to take the content on those sites down.
    – Mast
    Oct 6, 2022 at 20:33
  • 2
    @Mast: The redaction is a courtesy, not a guarantee. One should assume that anything sensitive posted to an Exchange site is already compromised.
    – Makoto
    Oct 6, 2022 at 20:39
  • 1
    @Makoto Oh, definitely. It's one of those voluntary responsibility things.
    – Mast
    Oct 6, 2022 at 20:56
  • 1
    API keys and passwords shared in the post are obvious. I'm talking about more unobvious things. For example, why does Facebook remove GPS metadata from uploaded photos? Do they have a responsibility to? Oct 8, 2022 at 4:07

2 Answers 2

13

No, likely for two reasons:

  1. Outside passers-by have no way of knowing if any piece of information is sensitive. It could be implied that it is (like for example, someone putting a DB password into a post - and yes it's happened more times than I care to count), but other data may not be so obvious. People put their email addresses into their profile, for God's sake.

  2. Stack Exchange assumes that you in posting on the site have the privilege to submit that information under CC-by-SA. If you're posting something that might be sensitive, you might not have the privilege to post it.

8
  • 1
    Ah, but if someone posts a PW in the question, that is obvious. But someone posting a link to a document that unobviously exposes their email address -- that is different. Oct 6, 2022 at 20:10
  • 2
    @IMTheNachoMan: It's obvious in that it's 60/40 in the certain that it's sensitive. For all I know it could be a dummy password. But I think that exposing an email address is less...how to say it..."critical" of a security issue than exposing the password to a production system. You can mitigate emails from bad actors, although it'll take a bit of effort. You've got a whole lot more work to do to fix production if your DB password gets out.
    – Makoto
    Oct 6, 2022 at 20:15
  • @IMTheNachoMan From a comment from your linked question "There has been an Update to Google's Sharing Policy on View Only Google Sheets. It no longer exposes the email address of the user. " It is the user's responsibility to share the document correctly. Oct 7, 2022 at 4:28
  • @DavidPostill If you read the rest of that comment chain, that was just a UI change. The information is still accessible through api even if you share "view only".
    – TheMaster
    Oct 7, 2022 at 8:32
  • 1
    @Makoto It's not about how critical the exposed data may be. It's about if the end-user knows they are exposing it. If an end user wants to remain anonymous then exposing their email address would be very bad for them. I work in cybersecurity so this is an interesting discussion. What responsibility does any site have in ensuring users/customers understand the unobvious data they are sharing. For example, why does Facebook redact GPS information from uploaded pictures? Oct 8, 2022 at 4:06
  • 1
    @IMTheNachoMan: I don't care much about Facebook since I don't use it. But honestly what you discuss speaks to reasonable OpSec for a responsible netizen. I have absolutely no way of knowing what it is you want to keep secret, so as long as you don't tell me, I won't know. But once you do tell me, then I presume that it isn't secret, since there's no presumption of trust or implicit guarantee of privacy on sites like this. And no, it is very expressly not Stack Exchange's responsibility to enforce good OpSec of its users.
    – Makoto
    Oct 9, 2022 at 3:37
  • @Makoto Why does SE have password requirements for the website? By doing so it is enforcing its users OpSec, right? Oct 10, 2022 at 3:48
  • @IMTheNachoMan: Password requirements have nothing to do with OpSec. There's an inherent benefit in using a password that meets a minimum length, though. OpSec is more a social nuance than a technical nuance, although it can be supplanted with technical help. But again, how you authenticate is so far divorced from what you publish that this herring is definitively bright neon red.
    – Makoto
    Oct 10, 2022 at 4:22
-2

While the question, whether it carries any responsibility, is speculative at best, I believe it's in the best interest of StackExchange, to do more to protect it's user privacy and as a byproduct, help itself in the process.

We have already established that sharing email address is not good for the community, even if it is what the user wanted, as it moves discussion away from SE to private. JorneymanGeek says

Likewise, an email, in addition to being a PII like, moves the information from SE and the commons to private messaging of sorts.

I'd consider this undesirable, as far as the goals of most sites go.

The case holds true for Google files in SE as well. Usually it moves the discussion away from SE to Google's inbuilt comments.

Yes, We are not 4 year olds, but accidents happen and there's a simple matter of ignorance. There are also precedents established by other Software companies like Github (and Facebook as mentioned by IMTheNachoMan). Github

Is Github used by 4 year olds? Why should they implement this feature? Because it helps users using it(and the companies, which provide the secrets)

In this case, it's trivial to detect whether there's a email address(or a Google docs link or password like characters) in a question and warn the user beforehand or explicitly disallow email addresses. The cost benefit ratio hugely favors implementing such a feature, especially with Stack Overflow.

8
  • Eh, I dunno. Unlike GitHub, you're not obligated to put your email address anywhere that's public, nor would it be tied to anything you do on the site. Similar to GitHub, you're in full control of what you publish to the site. GitHub's secret scanning tools or secret detection is a good-faith notice, not an attempt to prevent you from publishing that information, as any passing security practitioner would tell you that if something is published beyond its scope, it is safe to assume it is automatically compromised.
    – Makoto
    Oct 28, 2022 at 15:49
  • @Makoto secret detection is a good-faith notice I never said it's SE's responsibility. I'm just contrasting the potential benefits of introducing such a feature against the (trivial) costs: 1. Self contained/good quality posts 2. Valuable Discussion maintained within SE 3. User privacy
    – TheMaster
    Oct 28, 2022 at 15:59
  • 1
    The costs are most certainly not trivial. Do you know how easy it is to generate a legit looking email address? Why should Stack Exchange then have the power to decide if my dummy data is actually real data? What about passwords? God forbid I actually use "password" or "123456" to connect to my production database, but if I want to include that as a part of the configuration example, who is Stack Exchange to stand in the way of me doing that?!
    – Makoto
    Oct 28, 2022 at 19:06
  • Lastly, don't underestimate the cost of doing something like this. The only reason it'd be worth it is if it impacted Stack Exchange's business, too. GitHub, Google, Facebook and others have a lot of trust to lose if they mishandle sensitive information that we provide them. On Stack Exchange, you are not obligated to provide them with anything. So, what you put out there isn't impactful to their business since a rational actor would have no expectation of trust/secrecy for a question or answer posted on the network, and if that content did have something sensitive, then oh well.
    – Makoto
    Oct 28, 2022 at 19:07
  • @Makoto 1.You're not making sense to me, when you make this distinction On Stack Exchange, you are not obligated to provide them with anything.. Maybe you could better explain. Am I obligated to provide some kind of api secrets to Github? Am I obligated to provide some GPS data to Facebook, when I upload pictures? I don't think so. 2. Why should Stack Exchange then have the power to decide if my dummy data is actually real data? All I'm asking is a warning. "This information looks sensitive. Are you sure you want to post it?" That is trivial to implement with some basic regexes.
    – TheMaster
    Oct 28, 2022 at 19:51
  • Sure, I'll break this down for you, with the examples and services you cite. 1. It is embarrassingly, paper-bag-over-head-level embarrassingly common for people to commit secrets into their code repository. If all of a sudden you now move to GitHub, welp - you've just committed your secrets to GitHub. It's not OK if it's a private repo, either - GitHub (the company) ostensibly has access to it anyway, and now they know. Doesn't matter what they do with it. Oops, now you have to fix that and that's a whole ceremony depending on how deep those secrets go.
    – Makoto
    Oct 28, 2022 at 19:57
  • For Facebook, since I don't actually use the site, I'll go with another common denominator - pictures. You can strip metadata from pictures so people don't know where you were at when you took a picture. Even if you delete something in Facebook, it's still kept on Meta's servers. They've perfected the standard of soft deletion.
    – Makoto
    Oct 28, 2022 at 19:59
  • For Stack Overflow specifically, because the nature of the question is technical, you are more likely to run into a false positive trying to identify something sensitive. There's also still a lot of debate on if an email address is sensitive - heck, people put their email addresses in their profiles, so you can't just blanket say that an email is sensitive. And again, if I do pwgen -cn 32, you can't tell if the password I'm using as an example is sensitive or not. Why give me all of these alarms and false flags if nothing really comes of it?
    – Makoto
    Oct 28, 2022 at 20:01

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .