This seems really potentially dangerous, which the article hints at. Potential for entrapment and the software can't detect irony. How could this possibly go wrong?
| For example, if the suspect does not appear to be
| enticed into having a conversation, the software
| can appear offended or get more insistent.
This almost makes it sound like it was designed for entrapment. If someone doesn't want to talk to the chatbot, it will try and force a conversation? Really?
That's an interesting question. I wonder if that could be a usable defense. "What did I do wrong? I just thought I was talking to one of those bots, not a real person!"
But I think this paedophile hunting behavior around the world is going very absurdly far, and causing more damage than fixing things.
If you want to get real problematic people, go get those Haliburton guys kidnapping girls below 10 to sell in Africa (yes, that not only happened, it is still happening, the US government decided that their service is more worthy than the problems they create).
Chasing a few amount of people that chat to minors, or like lolicon, or like even movies of real people, is better to catch the wrong people than to do anything of good.
I took the time to read all of the first ten results and watch the video included amongst them.
There's two separate incidents, Halliburton subcontractor KBR and 12 nepali men, and DynCorp and Sex slaves in Bosnia.
Neither of which are Halliburton trafficking young girls to Africa, neither of which are any where near the continent.
Would you care to elaborate?
Its funny but it really would make for a fabulous Stross-ian sci-fi novel. What if a strong AI accidentally developed whos "prime directive" was to find pedophiliac tendencies in people and legally entrap them? Give it lesser god-like powers and you've got yourself a novel!
"Hanna, the CEO of Hofvarpnir Studios, just won the contract to write the official My Little Pony MMO. Hanna has built an A.I. Princess Celestia and given her one basic drive: to satisfy everybody's values through friendship and ponies. Princess Celestia will satisfy your values through friendship and ponies, and it will be completely consensual."
I'm glad you enjoyed. It is written by a LessWrongian, I believe. Also, it doesn't feel quite right to call either of these stories fan-fiction, as the word is usually correlated with mediocrity.
I thought Friendship is Optimal, although quite obviously suboptimal, wasn't that bad of a FOOM. For a much darker take on the subject, check out The Metamorphosis of Prime Intellect.
"Lawrence had ordained that Prime Intellect could not, through inaction, allow a human being to come to harm. But he had not realized how much harm his super-intelligent creation could perceive, or what kind of action might be necessary to prevent it.
Caroline has been pulled from her deathbed into a brave new immortal Paradise where she can have anything she wants, except the sense that her life has meaning.
Now these two souls are headed for a confrontation which will force them to weigh matters of life and death before a machine that can remake — or destroy — the entire Universe."
Actually I just meant that it's not unusual for 14 year olds to be generally attractive. Some teenagers are models and actresses and popular because of their mature looks. If you seek out relationships with 14 year olds, then that's some kind of paraphilia, but if a teenager (virtual or otherwise) starts making the moves on an adult man, then you just hope he has the willpower and sense to say no even though, biologically, you couldn't blame him for being interested.
The worst part is that 14, 15, 16 were very common marriage ages for women in our own civilization until a few generations ago. Since being physically attracted to girls that age is now considered a perversion, the inescapable conclusion is that the majority of the great people in history, and of our own ancestors, were pedophiles. We are all descended from pedophiles, and our civilization stands on the shoulders of giant pedophiles. That should make people more than a little uncomfortable.
I'm from Virginia, and as part of the University of Virginia tour they always show you the one-time dorm room of Edgar Allen Poe, famous American poet. During the tour, some snarky potential student will always point out that Poe was both a paedophile (he married his bride when he was 27 and she 13), and also a proponent of incest (she was his 1st cousin).
The quality of the research, and that's my own opinion, leaves very much to be desired (many fuzzy assumptions, a lot of hand waving & of the "and then, magic happens" syndrome). This wasn't published at an ACM or IEEE conference, which is typically a first warning sign.
Sadly, the BBC (and other major media outlets, but it's the 2nd time I observe it with the BBC website this month) has a knack for taking academic papers from a year ago (or more), and writing subpar articles about it.
Completely agree. Also the use of Google Translate is not going to be scalable at all. Many of the premises are theoretical and the criteria used to determine "level of disturbing content" is at best laughable.
"But researchers admit that it does have limitations and will need to be monitored. Although it is has broad conversational abilities, it is not yet sophisticated enough to detect certain human traits like irony"
Despicable. I'm still appalled at how hysterical the reaction is to this incredibly minor threat. Talking to a chatbot, it doesn't get more victimless a crime. This is just plain entrapment for a harmless thoughtcrime.
I once read that paedophiles have on average a lower IQ, I wonder if that is because they only catch the dumber ones. I know a chatbot wouldn't fool me. (And don't misquote me on this, I'm not admitting I'm a paedophile, you know what I mean.)
I feel ridiculous having to claim this up front, but I am also not a pedophile... however, I would very much like to find this chatbot and have a very long conversation. In fact, I might even be inclined to write a reverse chat bot that actually initiated conversations to this chat bot from a myriad of IP addresses, just to illustrate the innocuousness of it all.
My question is, what is the crime if someone is "chatting" with a bot? Any crime I could imagine would require a bad actor to actually engage with a person - how is it a crime if there is no victim?
Maybe I'm just naive or missing some apparent point but this really mystifies me.
The chatbot could be used to harmlessly engage with people in chatrooms, and flag 'grooming behaviour'. Then, those conversations could be stepped into by an actual person - saving having actual people monitoring the chatroom full time. The police move in when the IRL meeting takes place, and the suspected paedophile arrives full of evil intentions - at which point, the police have plenty of profiling experience and the guy probably takes a plea.
So the chatbot is more about whittling down the suspicious participant list, without soaking up police time.
It's not "victimless" when the same sexual predator engages with a real person and starts trying to obtain sexy pics and blackmail material with a view to meeting later down the line. I'm not a fan of the specific implementation of the bot described in the paper, but I'm quite happy with the basic principle that people trying to sexually corrupt algorithms should be in jail alongside people trying to hire imaginary hitmen or traffic imaginary slaves.
Besides, if chatrooms were widely suspected of being full of anti-paedophile chatbots, it would be a pretty good deterrent from paedophile participation even if they were actually pretty ineffectual at catching people. From what I remember of chatrooms from when I actually was about 14 - not many of the human participants would have passed the Turing test - the standard of conversation in a room full of bots might actually improve...
This. I don't recall the source or many details, but there was a guy who had chatted with a law enforcement official posing underage in a chat room, but he never showed up for their "date". They sent the SWAT team and news media to his house and he shot himself. They made the point that lots of people suspend belief to live a fantasy in a chatroom, knowing the other party is likely not who they say they are. Was this any different? Certainly a moral gray area, but not worth the guy's life for some sensationalism.
Like many others are saying, this does sure look like entrapment.
I cannot recommend the movie Outing [1] enough! It a documentary following a paedophile in Germany. Turns out there are a lot of "non-practicing" paedophiles that are recognizing that they shouldn't act on their desires, but they still have those desires. I think they'd be very easy targets for entrapment if someone approaches them like this.
"For example, if the suspect does not appear to be enticed into having a conversation, the software can appear offended or get more insistent."
This is a bit concerning, this feature sounds like it may try and provoke people into conversation to purposefully trap them, I really hope that's not what happens...
"But researchers admit that it does have limitations and will need to be monitored. Although it is has broad conversational abilities, it is not yet sophisticated enough to detect certain human traits like irony"
Luckily people using informal internet chatrooms never make ironic statements, so this software will be effective.
lol entrapment bot. maybe they can also release terrorist bot that denounces the west at set intervals and constantly private messages you "Blow up the embassy y/n? ... pls hold for FBI operator"
The conventional wisdom is that adults who solicit minors online typically use deceit (they have to, since minors have no interest in sex), and if you want to deceive someone into having sex in Spain, your victim has to be at least 16 years old. (Source: your link) So the chatbot would still catch criminal activity.
I can't believe so many people in these comments are apologists for paedophiles (and related terms). Having a problem with the way it is policed is okay. Accepting the act is not.
48 comments
[ 3.4 ms ] story [ 94.2 ms ] threadIt's a cool sign of the times that such questions can even be asked somewhat seriously and not just as a sci-fi future contemplation exercise.
But I think this paedophile hunting behavior around the world is going very absurdly far, and causing more damage than fixing things.
If you want to get real problematic people, go get those Haliburton guys kidnapping girls below 10 to sell in Africa (yes, that not only happened, it is still happening, the US government decided that their service is more worthy than the problems they create).
Chasing a few amount of people that chat to minors, or like lolicon, or like even movies of real people, is better to catch the wrong people than to do anything of good.
http://www.washingtonpost.com/wp-dyn/content/article/2008/08...
http://en.wikipedia.org/wiki/Halliburton
On those pages, Ctrl + F 'trafficking', 'rape'.
easy congratulations all around, more funding, happy PTA parents, "justice"
Future AIs will have a predilection for Bieber and meeting with strangers.
http://www.fimfiction.net/story/62074/Friendship-is-Optimal
"Hanna, the CEO of Hofvarpnir Studios, just won the contract to write the official My Little Pony MMO. Hanna has built an A.I. Princess Celestia and given her one basic drive: to satisfy everybody's values through friendship and ponies. Princess Celestia will satisfy your values through friendship and ponies, and it will be completely consensual."
I thought Friendship is Optimal, although quite obviously suboptimal, wasn't that bad of a FOOM. For a much darker take on the subject, check out The Metamorphosis of Prime Intellect.
Warning: Extreme triggers.
http://localroger.com/prime-intellect/
"Lawrence had ordained that Prime Intellect could not, through inaction, allow a human being to come to harm. But he had not realized how much harm his super-intelligent creation could perceive, or what kind of action might be necessary to prevent it.
Caroline has been pulled from her deathbed into a brave new immortal Paradise where she can have anything she wants, except the sense that her life has meaning.
Now these two souls are headed for a confrontation which will force them to weigh matters of life and death before a machine that can remake — or destroy — the entire Universe."
Unless, of course, the evaluator understands TXTinese.
Reminds me of the classic Monkey Dust sketch: https://www.youtube.com/watch?v=APG0rNedEwk
Something tells me they don't really know what "paedophile" means.
Direct link to PDF: http://paginaspersonales.deusto.es/claorden/publications/201...
The quality of the research, and that's my own opinion, leaves very much to be desired (many fuzzy assumptions, a lot of hand waving & of the "and then, magic happens" syndrome). This wasn't published at an ACM or IEEE conference, which is typically a first warning sign.
Sadly, the BBC (and other major media outlets, but it's the 2nd time I observe it with the BBC website this month) has a knack for taking academic papers from a year ago (or more), and writing subpar articles about it.
Talking to Negobot would probably be the equivalent of this: http://gizmodo.com/bank-of-americas-twitter-account-is-one-r...
http://dailycaller.com/2013/06/27/texas-teen-makes-violent-j...
Yeah, I don't think I want this program to be trying to entrap people then turn the police on them.
I once read that paedophiles have on average a lower IQ, I wonder if that is because they only catch the dumber ones. I know a chatbot wouldn't fool me. (And don't misquote me on this, I'm not admitting I'm a paedophile, you know what I mean.)
My question is, what is the crime if someone is "chatting" with a bot? Any crime I could imagine would require a bad actor to actually engage with a person - how is it a crime if there is no victim?
Maybe I'm just naive or missing some apparent point but this really mystifies me.
The chatbot could be used to harmlessly engage with people in chatrooms, and flag 'grooming behaviour'. Then, those conversations could be stepped into by an actual person - saving having actual people monitoring the chatroom full time. The police move in when the IRL meeting takes place, and the suspected paedophile arrives full of evil intentions - at which point, the police have plenty of profiling experience and the guy probably takes a plea.
So the chatbot is more about whittling down the suspicious participant list, without soaking up police time.
Besides, if chatrooms were widely suspected of being full of anti-paedophile chatbots, it would be a pretty good deterrent from paedophile participation even if they were actually pretty ineffectual at catching people. From what I remember of chatrooms from when I actually was about 14 - not many of the human participants would have passed the Turing test - the standard of conversation in a room full of bots might actually improve...
Also, and this is debatable, talking with a chat "robot" in no way constitutes the crime of pedophilia.
I cannot recommend the movie Outing [1] enough! It a documentary following a paedophile in Germany. Turns out there are a lot of "non-practicing" paedophiles that are recognizing that they shouldn't act on their desires, but they still have those desires. I think they'd be very easy targets for entrapment if someone approaches them like this.
[1] http://www.imdb.com/title/tt2076292/
Luckily people using informal internet chatrooms never make ironic statements, so this software will be effective.
http://en.wikipedia.org/wiki/Ages_of_consent_in_Europe#Spain