Chat Bot

Am I right to be concerned about the possibility of Chat Bots on this site?

There are a few concerning posts and, as I don't know much about these things, I'm not sure if I'm worrying over nothing here? Can they cause harm to individuals in any way on this type of chat site? 

If I'm being ridiculous, someone please tell me so! (I will be more relieved than offended.)

  • Concerning that last Post with the four quotes, there...

    Thanks, DongFeng5 for the reply. I tried about 5 times to edit this Post, to erase the Quotes to shorten it... but I cannot! Another thing to note, maybe: always leave spaces between multiple Quotes...?

    :-/

  • A New Thread has been started by Cassandro, and so I post a link to it here...:

    http://community.autism.org.uk/f/miscellaneous-and-chat/12621/copybot-and-forum-security

    Thanks very much... I am quite glad that I myself did not have to start it (!). I do not know when I shall write upon it, however. (I am currently going through a 'weirdness', nothing to do with this Forum.)

  • Never mind, it wouldn't work for long anyway. All the bot author has to do is add an extra space between two of the words in the pasted content, and the hash value will be something different, defeating the check. The NAS would have to look at this class of software instead:

    https://en.wikipedia.org/wiki/Plagiarism_detection

    That's a better fit for the problem at hand, and should give a score on how similar a new thread submission or reply is to an existing one that could have been automatically mined by some executable code.

  • 1. The NAS create their own daemon to peruse each thread OP and subsequent reply on the site, generating hashes of the bits which are NOT a quote of other users' content, then they store them in a SEPARATE database.
    2. When a new thread or reply is created, it initially goes into a holding area where the NAS automated tools can scrutinise it but the rest of us can't yet see it.
    3. An automated tool calculates a hash of the new content, then compares it against the existing database of hashes, resolving collisions as required.
    4. Any content where a duplicate is being submitted will be flagged for human moderator approval, alongside a copy of the original content which hashed to the same value.

    ...The "Quote"function does not always work correctly, and so I had to quote four times in order to re-quote all of that. I do so in the hope (...) that NAS (WebPM, probably) notices it and reads it.

  • Glad Tidings... I am still looking, and I see this. (I am not good at chat, though). Thanks... I am a bit calmer, now!

  • Suggested title: Potential security risks of using the NAS forum at home or one's place of employment.

  • Mr.Missile/CarCompany (!)... 

    I was going to type some replies to this Thread... but then you come here and fairly explode a grand amount of perspicacity. 

    I do not not know what to say...! But do you recall the ASCII Thread which I began  a while back? All of what you and Cassandro (and "Ellie") say... is fairly screaming out for this New Thread!! WebPM even once joined your good self in conversation, yet I interrupted it, suggesting a new Thread there as well - do you recall that?

    THIS THREAD IS LONG! NEW THREAD! NEW THREAD, PUHLEEEESE!

    (If I have to start it, then suggest titles, please...?)

  • In this sense, one may perhaps view an online autistic community as a fertile plain in which to grow a botnet capable of subsequent DoS attacks on multiple third parties.

    Large corporates are now more wary of phishing attacks and have trained their staff better accordingly, but -- oh hang on a minute -- none of you actually use this site from your place of work (or on the machine you use for work at home or banking), do you?!?!?

    If any of you answered, "Erm... yes...?", that's probably the explanation. Persons unknown have twigged that autistic folk may be at times vulnerable and naïve. That makes us easier pickings for phishing and -- worst of all -- a very "social-engineer-able way-in" to many companies which happen to employ autistic staff. Far easier to get someone autistic to click this harmless link when they are relying on these forums for support, rather than hope that unsolicited emails offering n0rp videos might still be successful despite the multi-level email filtering that is employed by most companies.

    Luckily -- so far, at least -- the fact that several of us (Disallowed Cynosure, Cassandro and others) have near-photographic memories is probably getting the better of many of the most ham-fisted attempts. Unfortunately, the authors will learn empirically from their failed attempts and evolve their unwelcome creations. And at least some of us may well have already been unwittingly pwned by the better attempts.

    Somehow, we shall need to collectively tilt the playing-field for them so that they leave us alone and go after softer targets instead...

    How about this lousy idea which won't work for long:

    1. The NAS create their own daemon to peruse each thread OP and subsequent reply on the site, generating hashes of the bits which are NOT a quote of other users' content, then they store them in a SEPARATE database.

    2. When a new thread or reply is created, it initially goes into a holding area where the NAS automated tools can scrutinise it but the rest of us can't yet see it.

    3. An automated tool calculates a hash of the new content, then compares it against the existing database of hashes, resolving collisions as required.

    4. Any content where a duplicate is being submitted will be flagged for human moderator approval, alongside a copy of the original content which hashed to the same value.

  • I suppose a CopyBot might have been deployed in the hope that it will manage to interactively regurgitate a formulaic "autistic journey"-type thread sufficiently well that it manages to earn enough trust from us to fall for it hook-line-and-sinker when it subsequently says click this harmless link...?

  • That reply with the picture was missing when I checked the thread. I checked everywhere several times and could not see.

    Now I could see.

    Maybe the missing replies problem?


  • Maybe we should see this all as water under the bridge, something already dealt with.

    Definitely water under the bridge, but more something to learn from and keep learning from in terms of respecting and protecting individuality.


    I think we all agree we want to play nicely here, but sometimes things will divide us.

    I am more inclined to communicating respectfully, for to play involves imagination at best and delusions at worst, and that is what divides the attention from the reality of what individuality involves.


    As I say, I could see why people might have thought R* was either (a) a bot; or (b) a scammer after personal details who might abuse vulnerable people on the forum. I could also see why people might think R* was (c) an individual with social and communication problems who should be afforded patience and people should try to get to know.

    Does anyone have any conclusive evidence for any of these?


    Yes, as presented above by myself and Cloudy Mountains on Chatbots and their operational programming. 


    The fact that the messages were short means there's not much data to go on. I think it's unlikely to be (a) because although there are chatbots that respond to natural language, I've not yet seen them used in forum spam; and also although the messages are repetitive, they weren't exact repetitions.

    Chatbots work by comparing what has been stated with the most likely reply for that given statement ~ using a dictionary of words with a thesaurus of phrases. Or in other words they function as written text calculators ~ adding and matching the coding of words with the coding of sentences.


    If the  know that R* is a genuine person (and not a Nexus-7 replicant), they could perhaps have reassured people that they have grounds for that (at the risk of speaking on behalf of R*).

    Ross and Kerri Mod already have ~ by stating the community rules which in essence obliges everyone to respect the individuality of all community members, which includes respecting also those who are new to or becoming familiar with this site ~ recalling this is a Public Domain with a world wide international audience of readers. There are those also who have joined this site but will not join in with posting because they have communication issues, a complete lack of self confidence, and actual written records across multiple threads that do not inspire as such confidence to start conversing. And, there are those who have left this site because of which, also.


    I don't think it's reasonable to suppose that rule 4 applies to spambots or abusers trying to get people's personal details for nefarious reasons.

    All rules apply to all registered community members without exception, and when the rules are not followed, whether this be as a result of an individual's indirect or direct registration, they are as such accountable for having broken the community rules.


    And given that we're not all great on reading between the lines, and there has been a problem with spam and spammers imitating humans, some degree of suspicion is understandable, surely?

    Chatbots converse as above described, and spambots do not converse but just basically leave a questionnaire that links to services, products, and or data storage systems.

    So as far as relating either of these to an individual person goes, it is not a reasonable hypothesis. Not knowing the community rules and or else not understanding the requisite information is understandable though, given the community rules have been broken, and the required comprehension on this subject is only really now starting to come together on account of your assessment. 


    I hope people here can understand where we each are coming from. I think the rules are sensible to avoid any legal liability and help people feel comfortable here.

    Most definitely ~ quote factor ten statement, instant classic the moment you posted it!


    The only purpose I can think of for continuing this thread is to reach consensus on how we might interact with R* in future.

    There is no need nor should there be any reason to reach a consensus upon how to treat any 'one' individual here at the NAS forums.

    It is a matter of individual governance and personal responsibility ~ i.e. the obligation of each individual to apply or as such develop the ability to respond respectfully and safely with all community members ~ accordingly as follows:


    1. Content must not be obscene, defamatory or libellous, vulgar, sexually orientated, hateful, threatening, or in violation of any laws.

    2. Content must not be sexist, homophobic, racist, intolerant of religious beliefs, or otherwise discriminatory. This includes links to porn or hate sites. 

    3. Content must not infringe any copyright, registered trade marks, rights of privacy, rights of confidentiality, publicity rights or other intellectual property rights, or other rights of any other person or organisation. 

    4. Content, including your story, message of support or profile entry, must not contain your, or any other Users’, personal contact information, such as: email addresses and personal website addresses, postal addresses, telephone, or fax numbers.

    5. If you are posting anything that could identify a person under the age of 15 you must either be that person’s parent or guardian or have the permission of the parent or guardian.

    6. Whilst mentions of products and brands are acceptable within your story or profiles, advertisements or promotions of specific products or services are not allowed and will be deleted.

    7. Links submitted in Content (as part of your story, message of support or in User profiles) must not link to web pages or services that break UK law or conflict directly or in spirit with the terms and conditions of use of the TogetherGifts site. Your account will be closed immediately if you link to illegal sites.

    8. Content must be posted in English.  


    The full community rules are accessible under Useful Links via the Terms and conditions link at the bottom left Arrow lower left of each thread page.  

  • I agree. I was thinking that it was like people on the forum had collectively formed an 'immune system' against spam and abuse. Although of course, as a mild hayfever sufferer, I know immune systems can overreact sometimes.

    Yours, TriviaBot

  • I am an “up vote bot”!

    Or like bot!

    I read, I agree, I respond, then I go hide waiting for the next awesome piece of kindness given or shown. 

    One thing we were all guilty of during the R debacle was,,,,,,,caring about each other, ok not always spot on, some panicked and did all they could to try and deal with it within thier abilities, others just warned others by saying it must be avaoided, some even tried to understand R and allow time to see if indeed it was a new user very much struggling to communicate, maybe a first time web user, unsure of social etiquette.

    regardless everyone was trying to do thier best to protect everyone else, 

    I like to think no one deliberately set out to be horrible, just try and protect.

    Wishing all a good day  , and much respect for our most enlightening new members, stick around I for one enjoy what you do and constantly learn new ideas and many new ways of thinking a thing through. You are awesome.

     Loving this community so much.

    x()x

  • None of the bots have passed the Turing test. Mitsuku and Cleverbot are capable of complex stuff but applying their code would be a very, very, long push to appear as a user who can converse with multiple users. The sheer amount of individual abstraction would be a throw out to any bot. It would take a long time. Those bots are directed towards a specific user usually to tailor a response that is specific to the user. It would need a fair bit of human direction to be any sort of threat to the users. 

    I know that there are spambots but that wasn't my main concern. Spam is a problem. My main concern was an actual human with bad intentions. This is something that is far more nefarious. I saw a few things going on that were targeting a user in particular. I thought that it was a bit strange. Bots don't profile a users personality quickly. BTW I've got to say you are *** hot at noticing spam! 

  • Maybe you're a ThankBot?

    ...I am a SocioPhobic-Winged-Cat -- partly Cyborg, maybe...

    A-a-anyway...(!) when I said "New Thread", I meant a new one about COPYBOTS... There are indeed so very very many types of SPAM, even here, but NAS is good at getting rid of them if they know about it (or if they pay attention)...

    WebPM asks for specific Threads to be begun about specific problems. That is what we are/were told, anyway, which is why I suggested it. WebPM is the one to pay best attention to, with regards to technical matters.

    Good Fortune to Yourself (and Thanks again!)  anyway...     :-) 

  • I cannot find the blueish-nighttime-cat themed picture..

    http://community.autism.org.uk/f/miscellaneous-and-chat/12186/new-friend-or-dating-rachael/69604#69604

    ...Here, that is a link to your own reply directly after it. Maybe press "Load Previous" and scan up or down to find the picture...? (?!)

    But this business is partly over with, now - Please heed most of all what DongFeng has Posted, here. (And Cassandro and Trainspotter also.) Please take care of yourself and stay careful and do not worry. 

  • I am Thanking a lot just at this point

    Maybe you're a ThankBot? Grinning

    Beside's DongFeng's ideas, I think 'Copybot' is somehow related to posting more obvious spam. Either it's trying to confuse anti-spam systems (by posting non-spammy text using a spammy method and spammy addresses that it wants 'cleanlisted'); or it's simply trying to increase number of incoming links to some spam sites (which it could add at a later point to its profile, or could edit the text people have replied to and replace it with spam links).  It's not clever enough to generate its own questions, so it parses some old ones out of existing pages.

    (New Thread, maybe...)

    We've some spam threads, and some threads about spam. Maybe even one or two threads about autism, if we're lucky.

    I suppose the thread we're in at the moment is partly about R*, but also chat bots generally.
    Here's the other about various spam postings and 'Copybot':
    community.autism.org.uk/.../mods-please-make-the-spam-stop

  • There remains something that I am very uneasy about regarding 'R'. And until a sensible post gets posted from her, I think we should just ignore any posts she makes, and then she will not concern any of us.

    Let's hope we get a post we can make more sense of.

    So R's approach for 'friends' (in the sense of connecting privately) only seems to have attracted three of us out of curiosity, rather than anyone genuinely wanting to connect.Not surprising given what you said.

    The relevant rules may be 2:

    This Community forum is public, so do not post personal or identifying details on it. This includes, but is not limited to, full names, addresses, email addresses and phone numbers.

    and 10:

    All the above rules apply to private messages, with the exception that people may wish to arrange to meet in person or contact one another by other means. In these cases please take the utmost care, and ensure that others are aware for your own protection. If you experience difficulties please contact the moderators for assistance.

    No one feels like it, so maybe she can get to know people 'in public' as it were.

  • Greetings to DongFeng5 (Missile/CarCompany!)...

    I am Thanking a lot just at this point, and you are one of them...

    Not everyone knows about ALL of this sort of thing, and so Thank You for Posting all of that --- ALL of it.

    (New Thread, maybe...)

  • Finally, yet by no means least... THANK YOU to TrainSpotter for understanding! I did follow your lead, in "friending" 'R', for reasons I stated before. I also posted link to the most "creepy"/strange part.

    Given your last paragraph, this was also my own conclusion... Yet when 'R' posts outside of their own Threads, then it is/was hard to try to say something. TrainSpotter, Thank You again for what you did and have said and still do.