2

I would like to know whats the evidence for this gotten closed?

Should we fear spammbots that write constructive and helpful posts?

I completly describe the concept which is intended to protect against bots. Thats somethign I have seen many times on here.

In addition I say lets assume captchas are outdated in this scenario, just to prevent answers stating "Captcha is ebtter because...." That got ignored and anyone was telling me anyway why captcha is better.

The I got complains about it beeing a "what if" kind of question. Thats kind of the case. But still just not because I wanted to create a hypotethical case but a question about "What wouldn't work with this one?" and not "Why other existing systems would work better?". But Actually I was receiving jsut answers for the latter.

And finally people agreed on the comment, that it is opinion based.

But I can't see why a question asking about failing points of such a system is opinion based. Can someone explain me? Or explain me whats else a reason for ti beeing off-topic?

| |
6

The Security.SE Help Center specifically says:

You should only ask practical, answerable questions based on actual problems that you face. Chatty, open-ended questions diminish the usefulness of our site and push other questions off the front page.

The idea of an AI which is so powerful to automatically solve CAPTCHAs and to automatically write useful comments based on a given context is not currently practical. While there are questions in this SE about the impact of quantum computers on cryptography, they are more practical because the technology is partially available and because basic ideas are already present (post-quantum algorithms).

Your question, on the other hand, takes for granted something that we cannot really understand, and that would bring so many complications that a definitive answer cannot exist (i.e. not answerable).

An AI with that level of sophistication could pass the Turing test, which means that it would be indistinguishable from human users. This, in turn, would mean that the spambot problem would be equivalent to the trolling problem on online communities, if the AI comments were not random but still not constructive, or to an even more complex problem, if the comments were also constructive.

For example, what would happen if human users rated as "not constructive" a comment which is actually meaningful, but conveys an opinion contrary to that of the majority of the users in the community? This is possible, and it already happens in some online communities. How could you distinguish between a constructive comment downvoted for this reason, and a comment downvoted because it is generated by an AI spambot?

Would it even make sense to talk about spam, and if so, how could we draw the line between this kind of "spammer" and a user with a different opinion from the majority?

This problem would not belong to information security anymore, not even remotely. Since you asked about an AI writing constructive comments, this is the only possible problem that would arise, and because it would not be an information security issue, your question is off-topic.

| |
  • All you are mentioning is nothing that my post tryed to catch. My post simply was about, "Given the system exists and works as a blackbox and there is actually the requirement for an alternate to captchas, what could make this concept fail, as the comic says, he actually developed it, so it exists." hamena314 got that point and his answer was pretty much the kind of answer I was looking for except, he still compared with captcha and why it can't compete to it. I just wanted to get a proval or disproval of this beeing widely usable not about its possible to be developed. – Zaibis Jun 8 '16 at 13:01
  • Admited this might be offtopic, depending on how "actual problem" is defined. but if that is the point, there are alot other questions which are obvisiously not based on actual problems. But probably my post by itself got mistaken, as it looks now to me what your answer here is tackling. – Zaibis Jun 8 '16 at 13:03
  • 1
    @Zaibis Even if you just want to know if this would work, the question is still open-ended, because it's not clear what the actual problem would be (see above: how would you define spam?). You commented Philipp's post by assuming that this system could use constructive answers from other websites in an appropriate context. If that was the case, I don't really understand what the problem you're trying to solve would be. Moreover, if this system existed, it would pass the Turing test, and thus any kind of (automatic or manual) test would fail to recognize it as non-human. – A. Darwin Jun 8 '16 at 13:20
2

Then I got complains about it being a "what if" kind of question. That's kind of the case. But still just not because I wanted to create a hypothetical case but a question about "What wouldn't work with this one?" [...]


"Given the system exists and works as a blackbox and there is actually the requirement for an alternate to captchas, what could make this concept fail [...]

Hypotheticals aren't always 100% off-topic, but to me it seems your question was definitely over the line of "too hypothetical to answer". When it comes down to it, SE is supposed to be Q&A site, not a discussion site. Still, there are definitely on-topic "discussion" questions on here, but they generally revolve around something practical and current, and even then it's a coin toss to see if it'll get closed as opinion-based.

Your question though, comes across to me as definitely too hypothetical to be answered in a concrete and factual sense.

| |
  • Its hypotethical, yeah but I don't get how there is even potential to get into discussion as long no one discusses about why my hypotethical case is not possible. Or is it even offtopic. If op says "Lets assume X is given[...]" and then everyone starts discussing about denying to assume X as given? But ok, I'll should just let it go I guess. – Zaibis Jun 8 '16 at 13:36
  • 1
    @Zaibis I think your issue here is that (usually) "Let's assume X is given [...]" is already out of scope for SecSE. The site exists to try and give helpful answers about what we know; asking a question that has to have some form of currently impossible setup in order to answer it, cannot realistically have a real or helpful answer. – WorseDoughnut Jun 8 '16 at 13:40

You must log in to answer this question.

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