This article http://www.newscientist.com/news/news.jsp?id=ns99994783 tells about a IT consultant that has released 100,000 bots in chat rooms world wide with the soul intent of catching pedophiles. The bots will chat with unsuspecting users, and if the bot thinks that the conversation is a preclude to “grooming”, it sends the transcript of the conversation to the consultant’s office where he can decide to contact the local authorities.
My first question is, are chat rooms littered with pedophiles? Is this really a problem, or are we just scaring ourselves into all of this effort?
My second question is, does this step over the line privacy of the chatters?
First, my opinion on the grandness of the pedophile problem is that I really don’t know if a problem really does exist. I rarely visit boards or rooms where the teen/child population is large, but out of the few times I have, I have never noticed any such activity. Does anyone have stats on this sort of crime?
Secondly, I really do have privacy issues over this. I am of the school of thought that states innocent until proven guilty. While suspicious activity should be reported, I do think that individual crusaders should not take it upon themselves to police the world. To me, the boarders on vigilantism. Not to mention the slippery slope argument of using this technology to weed out other types of conversations.
This isn’t an electronic neighborhood watch. That would simply be monitoring public chat rooms for suspicious activity. This is a sting operation. It could quickly stomp on the rights of the accused since the accusers have no oversight, liability or training in criminology (what is a “prelude to grooming”? How does an AI recognize it?) or the law.
I have no idea if Internet pedophiles are a real problem or not, but if they are, sting operations (as opposed to minitoring) should be conducted by legal authorities.
Since the bots have no authority to arrest anyone, and apparently do nothing more than notify authorities of suspicious behavior, I don’t think the term “sting” applies. Granted, the “engagement” of potential pedophiles is a little more involved than just watching a suspicious person in the neighborhood, but not by much.
Neighborhood watch: “Hello, police? I was out walking my dog and chatted briefly with a suspicious-looking person wandering in the neighborhood. Can you check it out?”
Pedophile bot watch: “Hello, police? I was out blabbing away mindlessly and talked briefly with a suspicious-sounding person in the KIDFUN chat room. Can you check it out?”
In both cases, the decision of whether or not there’s any illegal activity remains with the authorities. If the bots don’t generate “false positives” too easily, I don’t think there’s a problem here.
I hadn’t thought there were any AI jobs in law enforcement, but it seems I was wrong. Of course, this guy isn’t law enforcement, but if he were it’d be a really cool idea.
If that log was really produced by a machine, it’s an order of magnitude beyond any AI chat programs I’ve seen. I wish he’d open-source it, but judging by his comments in the article he’s not that kind of person.
(Well, I suppose you could. You could have claimed that amateur sting operations caused World War II, or amateur sting operations are responsible for entropy in the universe. So forget the hyperbole. Let’s just say you’re wrong.)
In general, there are no legally cognizeable rights that are being trod upon by this process. No one has a “right to interact in chat rooms without wrongly being reported as a suspicious person,” for example. As long as private citizens, and not government agents, are doing the monitoring, their conduct is unlikely to infringe upon “rights.”
Now, this may not be the wisest approach to the problem of curbing Internet pedophile “grooming.” The number of false postives reported may put a strain on law enforcement respurces, hindering efforts to catch real pedophiles. But there is no problem of “rights” here.
To get an idea of how much pedophilic (is that a word? heh) behavior goes on in chatrooms, and what some people are doing about it, check out http://www.perverted-justice.com
And on that very site you can see just how far some of the ‘baiters’ are willing to go in order to keep the person they call a pedophile, and whose life they attempt to ruin through harassment, intimidation and blackmail, talking to the supposed ‘child’ about sex.
If some of those chats were conducted by law enforcement, they’d very likely be thrown out as entrapment, at least in my non-lawyer opinion.
Saw a show on TV where a TV station did some baiting out of Detroit where the perps came to what they thought was the house of a 14 year old girl for some sex play. 15 showed up and they got them all on camera. Pretty scary! I think that if you are an adult playing in kid oriented chatrooms, then you deserve to be tripped up. There is no good reason for you to be there.
The “bot” part of the conversation in the New Scientist link not only responded to the chatter, but brought up new subjects. I can easily imagine that the bots already talk to each other (without realizing it) at least from time to time. I would like to know if one of these bots has ever reported another bot for “grooming”. I am not saying that this would prove anything (except, perhaps, the impartiality of the reporting), but it is an interesting idea.
For what it is worth, the fact that I couldn’t tell which part of that conversation was synthesized (or, rather, which was authentic) shows me once again why I never chat, or even post in forums outside SDMB.
First, there’s a long history (relative to Internet Time) of IT consultants posing as Internet vigilante do-gooders to get publicity. See here. Reporters who are technically less literate than programmers are easily fooled into thinking there’s a story there.
Second, natural language processing, of the type implied by this article, just isn’t there yet. If it was, it would be a revolution in the field of artificial intelligence. What he may have (if he actually has a bot), is Eliza jacked up on steroids with a big dictionary and a lot of sentence patterns.
Third, notice that the IT consultant claims to have been responsible for starting police investigations, but no specifics were mentioned, and the magazine carefully distanced itself from the claim by reporting that they couldn’t verify any of them.
So what happens when some IT type puts 200,000 bots in different rooms to mimic a pedophile, as a protest of some sort? Would be interesting wouldn’t it?
I used to frequent their site months ago, and I remember at least one guy who kept trying to steer the conversation away from sex, but the ‘kid’ kept steering it back, but I don’t remember what that nick was, and I don’t have the kind of time to go through every log to find it.
Although you can get a glimpse of another huge problem I have with their site, their practices, and their mentality if you register for and lurk in their forums. The outright glee they feel at ruining someone’s life without a trial of any kind is every bit as disturbing as pedophilia.
I raised these issues the last time this topic was discussed here.
Searching the web, I found these words of one of their former members:
The veracity of that accusation is no more or less determined than the veracity of accusations perverted justice makes and attempts to ruin lives with. This remark came in response to the PJ ‘outing’ of a Michigan Teacher named Brian Graves, who then attempted suicide.
Just to throw an idea out for general chewing, what exactly is the problem if a 14 yo girl invites an older man round for “sex play”?
There is nothing over and above the kinds of concerns that would be applied to anyone, of any age, that did such a thing, yet there are not “10 bazillion” bots out searching for anyone who uses the net for picking up lonely housewives, for example.
Is that we have the perception that a 14yo is not able to make decisions properly, or is it that we do not think she is ready for sexual activity?
Are people not allowed to make their own mistakes?
If this hypothetical girl is up to the point of initiating sexual encounters, is that not a sign that perhaps she is ready for sexual activity?
Anyway, a pedophile is defined as one who is sexually interested in prepubesants, not 14 yo per se, so in this case the term is applied incorrectly.
Perhaps a sign of how this is, as the OP says, being used to create another layer of fear in this world.
If the bot in the original article is true, then it is pretty scary. I couldn’t tell it from the real person. You could all be bots for all I know, I know, I know, I know… sorry oil pressure dropped a little there.
Those guys at perverted justice are a bit over the top though. While I am quite happy them busting the perverts in the chat rooms, do they have to run a poll to find the most perverted conversations?
I have read perhaps a dozen of the chats posted on PJ. I have never seen anything that would, if done by a law enforcement officer, qualify as entrapment. With your admission that you can’t point to any such conversation either, I think we can safely table the accusations of ‘entrapment.’
Mr. Graves’ case does not create any sympathy in me, except, of course, for the natural sympathy for anyone considering suicide. But if Mr. Graves had been arrested as part of a law enforcement peration, he equally well might have tried to commit suicide.
His conduct was wrong, and being exposed is a foreeable to just risk that someone engaging in his conduct takes.
What disturbs me, however, is this paragraph in the story:
I have a great deal of sympathy for that man, and believe he should take legal action against the site. Graves has little to legitimately complain about; his hands are not clean. But the wrongly accused teacher has a legitimate cause of action.
I’m sorry, I can’t remember any nicks from five or six months ago that seemed to me to be entrapment. I do remember that one of them was a clergy member who steered the conversation away from sex as much as possible, and then went on to say that he counsels troubled teenagers as part of his job. Another person they accused was a man who told them that he was out of town at the time they claimed he was chatting up their fake ‘kid’ and that his own teenage nephew had access to his computer during that time.
They, the ‘perverted justice’ people, still insisted on calling both of these people, their families, their employers, and putting up fliers around their neighborhoods to harass and intimidate the alleged ‘guilty’.
I also have a severe problem with the logs they post because a log is no more than a text file, and it’s very, very easy to go back into one after they’ve already chatted with someone and change information. There is no proof whatsoever that these logs are 100% legitimate, that they haven’t been altered by the people at perverted-justice.
This is especially a concern after having read what one of their former baiters had to say about their techniques of ‘scouting’ their potential marks and initiating sex talk, and of course omitting all of that from the ‘log’ they tell us to accept at face value, and upon which basis they want someone’s life ruined.
These people are not law enforcement, Bricker, and what they’re doing sounds to me a lot like blackmail. Through unverifiable methods they claim they got a guy to talk about sex to an ‘underage’ person, and then they turn that person’s real name over to others who, unless the accused does exactly what perverted-justice wants, will attempt to get that person fired, divorced, abandoned, and otherwise ruined.
As to the first example - so what? What you described doesn’t reach the level of entrapment. If, despite his initial demurs, the clergyman arranged to meet an underage partner for sex, then neither the law nor I have any sympathy.
As to the second example - it has nothing to do with entrapment. The second man’s defense was not that he was enticed into committing a crime that he otherwise would not have committed, but rather that he was in fact innocent: he never did it. I’ve already acknowledged, based on the misidentification of the first teacher in the Brian Graves story, that false accusations (as opposed to entrapment) are a serious concern with this process.
Interestingly, none of the “Right-of-Reply” messages I’ve seen suggests that the chat logs have been altered in any way. It seems to me that if this had happened, at least one of the people unfairly smeared by the altered log would be making that accusation. Do you have any evidence that this has happened, or are you just voicing concern about the potential?
If PJ were demanding money, then I’d be outraged. But what is it they actually demand of the people they “catch?”