There is a constant war between spammers and website operators, to prevent websites from spam. CAPTCHA's are the tools to protect sites, and are the front line of this arms race.

This is an area of AI research that is directly relevant to the public.

The question is if we should allow postings about how to defeat CAPTCHA's. They are probably in scope, but we don't want to help spammers.

Obligatory XKCD link.

We may get some inspiration from Security.SE. They have some experience in dealing with ethical issues. Over there, they have an explicit close reason for questions about hacking other systems:

Questions asking us to break the security of a specific system for you are off-topic unless they demonstrate an understanding of the concepts involved and clearly identify a specific problem.

Maybe we need a similar close reason or off-topic flag?

  • $\begingroup$ Is that Security.SE quote really the same thing? The way I understand it is that it is mainly meant to stop questions like "I want to hack example.com. Help me!". Aside from that, there's plenty of "abusable" posts over there... $\endgroup$ Aug 5 '16 at 11:30
  • $\begingroup$ As for the CAPTCHA's, wouldn't any descriptions which were detailed enough to be useful to spammers be off-topic anyway? I mean, just saying what the general approach is doesn't really help anyone. $\endgroup$ Aug 5 '16 at 11:32
  • $\begingroup$ @InquisitiveLurker It is one example of how they deal with possibly unethical issues. Note that it is not explicitly ruling out "black hat" activity. I think we can take some inspiration from them - from this close reason and other ethical issues they've dealt with. Even if we decide that we are not going to follow their lead, at least we'll have considered why not. $\endgroup$
    – S.L. Barth
    Aug 5 '16 at 11:52

(Without seeing any actual examples…)

These can likely be closed as off topic already. A question sufficiently detailed enough to ask how to defeat a system based in AI will likely no longer be about the subject of AI itself. It would be like asking how to defeat the smart aliens in Galactic Uberblast 2020, or how to remove the T47/a access panel of your robot butler if he wont let you.

There's a point where a question is only coincidentally related to the subject of AI itself — close it as off topic.

There will always be sticky edge cases where someone might be asking how an AI-based security system works, but it becomes somewhat problematic to preempt any such questions by presuming the intent of an author before you see such questions in actual practice. It's probably too early to conjure up a broad policy statement when there is really no tangible problem to defeat.


Asking about how Captcha works or what is mechanism of AI recognising the text from the image, isn't illegal.

Neither whether has been cracked/hacked or not, which has been asked at Stack Overflow:

Captcha is just a type of challenge-response test used in determine whether somebody is human or not and this technology is not owned by anybody. There are many research studies how to improve this technology to keep spammers away.

If the post doesn't indicate it's against the law and doesn't show any illegal activity, there is no reason to close it. If it's unethical, you can always down-vote it or ask for clarification. There are special agencies which deals with that problem, so you don't have to worry about it.

If it's not illegal, it's up to you how you'll use the knowledge from the posts. As it can be always used for research and educational purposes.

See also:


You must log in to answer this question.

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