Beta
×

Welcome to the Slashdot Beta site -- learn more here. Use the link in the footer or click here to return to the Classic version of Slashdot.

Thank you!

Before you choose to head back to the Classic look of the site, we'd appreciate it if you share your thoughts on the Beta; your feedback is what drives our ongoing development.

Beta is different and we value you taking the time to try it out. Please take a look at the changes we've made in Beta and  learn more about it. Thanks for reading, and for making the site better!

Tracking Pedophiles By Their Typing Habits

CmdrTaco posted more than 3 years ago | from the you-gotta-be-kidding dept.

Science 292

An anonymous reader writes "Pedophiles using the internet to target youngsters could be tracked down — by the way they use a keyboard. Researchers are investigating ways to use technology which can determine a typist's age, sex and culture within ten keystrokes by monitoring their speed and rhythm." Since Phrenology hasn't exactly panned out, they gotta try something new.

Sorry! There are no comments related to the filter you selected.

oh no (5, Funny)

genican1 (1150855) | more than 3 years ago | (#31609524)

i hope they don't mistakenly target me in my rush to type fast and make this the first post!

Re:oh no (5, Funny)

K. S. Kyosuke (729550) | more than 3 years ago | (#31610082)

They just might; after all, you seem to be excessively aroused by underage articles.

Re:oh no (4, Insightful)

Jurily (900488) | more than 3 years ago | (#31610250)

How does this work anyway? Does it account for the fact that I know my keyboard is wearing out and thus tend to hit the keys a bit stronger? Does that make me a man?

Does it account for the fact that I grew up typing, and don't even need to look at the keyboard or screen while typing, because it "flows"? Does that make me a woman?

Does it distinguish between typing on a laptop keyboard and a regular one? Since I'm used to laptop keys, I tend to go softer on regular keyboards. Does that make me a pedophile?

And most importantly: WHAT DOES MY TYPING SPEED HAVE TO DO WITH WHO I WANT TO FUCK?

Foolproof (4, Insightful)

MyLongNickName (822545) | more than 3 years ago | (#31609526)

I am sure this is foolproof! Sign me up!

Re:Foolproof (4, Funny)

somersault (912633) | more than 3 years ago | (#31609644)

Congratulations - you are a pedophile! Either that or a software engineer, but we've lost your paperwork and are just going to go with pedophile to save some time.

Re:Foolproof (4, Funny)

mrsteveman1 (1010381) | more than 3 years ago | (#31609784)

Congratulations - you are a pedophile! Either that or a software engineer, but we've lost your paperwork and are just going to go with pedophile to save some time.

Apparently you've never seen Japanese video games, there's no difference.

actually it will work (5, Funny)

Anonymous Coward | more than 3 years ago | (#31609530)

they're hunt-and-peckers.

Re:actually it will work (0, Redundant)

Pojut (1027544) | more than 3 years ago | (#31610318)

Pecker? I barely even know her!

Left mousing? (5, Funny)

ThisIsAnonymous (1146121) | more than 3 years ago | (#31609544)

Looking for "left mousing" would be more accurate...

Re:Left mousing? (0)

Anonymous Coward | more than 3 years ago | (#31609956)

I masturbate with my left hand, you insensitive clod!

Re:Left mousing? (1)

PhilHibbs (4537) | more than 3 years ago | (#31610052)

I use the mouse left-handed at work, but right-handed at home for gaming.

Re:Left mousing? (0)

Anonymous Coward | more than 3 years ago | (#31610130)

I use the mouse left-handed at work, but right-handed at home for gaming.

So you only molest kids when at work?

Typical /. summary (4, Insightful)

wowbagger (69688) | more than 3 years ago | (#31609566)

Typical /. summary - conflating the idea of "we can tell how old you are from your typing" with "we can tell if you are Wicked Uncle Ernie by your typing."

The idea of this would be that if you can tell that somebody is a 40 year old man, and they are on a forum saying they are a 12 year old girl, you can flag that as being suspicious.

Of course, the problem is that a web site or chat forum has relatively little visibility of the user's typing pattern, so unless you force all forums to be accessed by special software that can monitor typing (AND you prevent the use of cut-and-paste so that you type in one window, then paste into the chat window, or detect such matters and flag THAT) then this won't be very useful at all.

Of course, making THAT statement in the summary would be hard, and would require actually thinking about the story.

As I said, typical /. summary.

And... (2, Insightful)

Frosty Piss (770223) | more than 3 years ago | (#31609736)

But even more, how does it follow that the highest use of research into this sort of "knowledge" will / should automatically be use to find pedophiles? Isn't this just more of the THINK OF THE CHILDREN *hysteria*?

Dvorak (4, Funny)

langelgjm (860756) | more than 3 years ago | (#31609794)

Joke's on them - I use Dvorak!

Of course, by saying that and posting on /., I've probably narrowed down my age, sex, and culture enough already. And I've identified my residence as parents' basement. Have they accounted for the mediating effects of Cheetos consumption on typing habits?

Re:Dvorak (4, Funny)

GigG (887839) | more than 3 years ago | (#31610146)

Joke's on them - I use Dvorak!

I think that pretty much guarantees we will be seeing you on 20/20 some day.

How It Does It (1)

Hoi Polloi (522990) | more than 3 years ago | (#31610262)

It watches to see if you are typing with just one hand.

Re:Dvorak (0)

Anonymous Coward | more than 3 years ago | (#31610266)

Dvorak? We already know such keyboard users are sociopaths, why would we need to know if they are _also_ pedophiles? One creepy personality trait per keyboard user is sufficient.

Re:Typical /. summary (2, Insightful)

MtHuurne (602934) | more than 3 years ago | (#31609840)

It's not hard to monitor typing into an edit box using JavaScript. The script could compute some kind of typing signature and then send that to the server. Copy-pasting all text is a very odd way of typing and could be flagged.

I doubt such a detection algorithm would be accurate for all people, so they should not start auto-suspending accounts based on this. But it would help if a human moderator gets some hints about which accounts to pay closer attention to, since manually monitoring everything is not feasible.

In the real world, we have Republicans. (1, Insightful)

Anonymous Coward | more than 3 years ago | (#31609848)

Please keep in mind that, in reality, we have to deal with Republicans, "Teabaggers", and others of their ilk.

They don't realize the limitations of technology. They don't particularly care about fair justice, either.

So they have absolutely no qualms about referring to somebody they dislike as a "pedophile", and then backing it up with this sort of keystroke analysis as "proof".

If you don't believe me, just look at how they justified the Invasion of Iraq. They showed computer-generated images of "weapons labs" and pictures of heating ducts, all in front of the UN!

Re:In the real world, we have Republicans. (0, Flamebait)

DarkKnightRadick (268025) | more than 3 years ago | (#31609958)

Please keep in mind that, in reality, we have to deal with Democrats, "Liberals", and others of their ilk.

Fixed that for ya.

Re:In the real world, we have Republicans. (0)

Anonymous Coward | more than 3 years ago | (#31610104)

Please keep in mind that, in reality, we have to deal with Republicans, "Teabaggers", and others of their ilk.

Fixed that for ya.

I have fixed that for you.

Re:In the real world, we have Republicans. (2)

L4t3r4lu5 (1216702) | more than 3 years ago | (#31610154)

Please keep in mind that, in reality, we have to deal with Muggles, "Mud-bloods", and others of their ilk.

FTFY

Re:In the real world, we have Republicans. (1)

DarkKnightRadick (268025) | more than 3 years ago | (#31610222)

lol

Re:In the real world, we have Republicans. (0)

Anonymous Coward | more than 3 years ago | (#31610320)

Please keep in mind that, in reality, we have to deal with Republicans, "Teabaggers", and others of their ilk.

They don't realize the limitations of technology. They don't particularly care about fair justice, either.

So they have absolutely no qualms about referring to somebody they dislike as a "pedophile", and then backing it up with this sort of keystroke analysis as "proof".

Given that the left-wing Labour party has been in power in the UK for many years, and has removed far more rights from its citizens than the Republicans ever did, I find your political labeling analysis to be pointless.

And President Hopenchange is no better.

Re:Typical /. summary (1)

Jahava (946858) | more than 3 years ago | (#31609890)

Of course, the problem is that a web site or chat forum has relatively little visibility of the user's typing pattern, so unless you force all forums to be accessed by special software that can monitor typing (AND you prevent the use of cut-and-paste so that you type in one window, then paste into the chat window, or detect such matters and flag THAT) then this won't be very useful at all.

It can certainly be implemented via a Javascript-hooked message box and periodic AJAX messages. Look at how Google implements chat in GMail or Wave. Hell, in Wave they can show you typing letter-by-letter. Granted there is some latency and blocking, but heuristics could certainly work to smooth that out. Data may be lossy, but it also may not be so lossy that nothing useful can be extracted.

There will always be ways around (Javascript blocking, copy/pasting, etc.), but your average Internet forum poser will probably not be aware enough to take those measures, especially if they result in user interface inconvenience.

Re:Typical /. summary (1)

ZeroSumHappiness (1710320) | more than 3 years ago | (#31610272)

Simple solution: Write a program that disguises your typing pattern as a 12-year-old boy. You can embed this in a keyboard if you really want to disconnect it from the computer. All it takes is some learning about what pattern you're trying to emulate.

Re:Typical /. summary (1)

fearlezz (594718) | more than 3 years ago | (#31609912)

Not gonna work. Most of my 40-year old collegues type like 12-year olds....

Re:Typical /. summary (1)

Lumpy (12016) | more than 3 years ago | (#31609978)

Yup.

so simply slap in a lot of OMG! and completely butcher spelling and grammar to thwart the "scanners"

Honestly, I could write a greasemonkey script to emulate the atrocious communications that teens and pre teens have. Hell that same script could simply act as one with some really simple AI rules.

OMG! Rlly?
u suk.
OMG!
OMG?
cl m l8ter k?
y
u suk!! :)
OMG!

Ok I feel like I just lost 30 IQ points...

Re:Typical /. summary (3, Interesting)

English French Man (1220122) | more than 3 years ago | (#31610156)

Problem is, there is nothing illegal in pretending to be a 12 year old girl when actually a 40 year old man. Or copy-pasting into textboxes to avoid typing pattern detection.

So we are talking about a method for detecting something that isn't illegal by itself (it becomes illegal when the 40 year old guy invites her new friend in his house to mistreat her) and flagging it as suspicious. Automatic crime detection leads to an unfair justice system IMHO. In addition, flagging someone for a crime that, statistically, they have more chance of committing than others I find wrong too.

Last, but not least, how many false positives? How many 12 year old girls type like 40 year old guys? What would be the consequences?

Re:Typical /. summary (1)

Kjella (173770) | more than 3 years ago | (#31610330)

The idea of this would be that if you can tell that somebody is a 40 year old man, and they are on a forum saying they are a 12 year old girl, you can flag that as being suspicious.

Somehow this sounds more like a tool for pedos to find out who's really a 12 year old girl and not a 40 year old man... Police sting? Anti-pedo vigilante? People just yanking your chain pretending? Aha... a real 12yo girl, jackpot. Not to mention I can imagine many really weird situations from false positives. There's so many holes it this logic there's more holes than logic.

Congrats! (3, Insightful)

boneglorious (718907) | more than 3 years ago | (#31609572)

Now good luck getting this pedophiles to have those sensors hooked up and letting themselves be videoed and monitored. How about saving yourself the analysis and just looking on their screen for child porn??

Re:Congrats! (1)

MyLongNickName (822545) | more than 3 years ago | (#31609626)

If it is just speed an rythm, wouldn't you be able to tell this in virtually any chat room?

No, I didn't RTFA. There may be nuances to this left out of the summary.

Re:Congrats! (1)

boneglorious (718907) | more than 3 years ago | (#31609660)

No, your chat room doesn't know I've typed until I've clicked "post". And even if it does, this probably requires more accuracy than you could remotely.

Re:Congrats! (1)

Rallias Ubernerd (1760460) | more than 3 years ago | (#31609694)

actually, they know your typing as soon as they send you the code for the text box to type. Windows live knows your typing immediatly.

Re:Congrats! (1)

clone53421 (1310749) | more than 3 years ago | (#31610038)

Yes, but they’d have to inject extra code because “virtually any chat room” doesn’t monitor your keystrokes like they’d require for this to work. Even if it says “typing” / “entered text”, that’s not enough.

Re:Congrats! (1)

Jesus_666 (702802) | more than 3 years ago | (#31610336)

Chatrooms and instant messengers are two different things. Yes, IMs could possibly monitor this, although it would be easily circumvented by using a multi-messenger which doesn't implement this or which reports bogus typing data. Those would be around simply because open source geeks like their privacy.

Chatrooms... Not so much. They're usually IRC with or without a frontend. While the frontend could monitor the typing habits, you can easily circumvent that by connecting directly. IRC is line-based so without a special client you won't get any data. Of course you can try to kick people off based on client etc. but you can bet that such measures will be cirumvented.
There's non-IRC chat, of course, but I'm not certain how popular and/or easy to get into it is.


However, that's all irrelevant anyway. This technology unlikely to be widely implemented in the first place. Imagine you were starting a website for kids and hadn't read this article. Would you think "hey, I need to investigate technical means to determine whether one of the kids in the chatroom is actually an adult"? Chances are you wouldn't. I'd wager that most operators of webites for children are less technically inclined than the average Slashdotter so they're even less likely to think of it. For every site with a chat that implements this and puts effective barriers in place to keep people from getting around it there's ten that don't even have an operator in half the channels.

Re:Congrats! (1)

MyLongNickName (822545) | more than 3 years ago | (#31609666)

Crap. Let me rephrase:

"with virtually any chat room software? I don't see the need for special equipment.

Re:Congrats! (1)

boneglorious (718907) | more than 3 years ago | (#31609680)

Remember, when I press a key, the computer doesn't register it immediately, unless I'm using special software that prevents it from being buffered.

Round trip (1)

wjousts (1529427) | more than 3 years ago | (#31609742)

More to the point, if the software that is supposed to analyze your typing patterns is on the server (and it would have to be, who would let somebody install this locally on their computer?) then they'd have to change the way their chat application works so that every keystroke is sent to the server instead of just sending the whole message when somebody hits "send". Can you say lag?

Re:Round trip (1)

BarryJacobsen (526926) | more than 3 years ago | (#31609786)

More to the point, if the software that is supposed to analyze your typing patterns is on the server (and it would have to be, who would let somebody install this locally on their computer?) then they'd have to change the way their chat application works so that every keystroke is sent to the server instead of just sending the whole message when somebody hits "send". Can you say lag?

Or the java/flash applet that presents the client interface can send some extra data about speed and rhythm after you hit post....

Re:Round trip (1)

English French Man (1220122) | more than 3 years ago | (#31610278)

This can be easily modified, and the hacker paedophiles would instantly workaround that.

Wait... what? (3, Insightful)

Kell Bengal (711123) | more than 3 years ago | (#31609576)

It boggles my mind that they even consider this a legitimate avenue of inquiry... I mean, really. How many opportunities do kiddy-fiddlers present for you to analyse their keystrokes, but not say... their IP? Sure, it allows you to tie an online interaction to a particular user, but if you have that kind of hook into their system odds are you know more than enough to do that already.

Oh wait, you mean there are applications where you want to tie users to content generically on the internet, but you think it's too invasive to make everyone agree to it simply so you can sell them more soda or send them to prison for their political ideas? That's OK - we'll just say it's targeting pedophiles.

That always works.

It's worse than junk (4, Insightful)

Joce640k (829181) | more than 3 years ago | (#31609706)

I'm pretty sure chat programs send the message in chunks, not individual keystrokes. /Didn't dignify the article with a reading.

Re:It's worse than junk (2, Insightful)

boneglorious (718907) | more than 3 years ago | (#31609760)

Exactly, and add to that the variances in how quickly things move over the internet, and suddenly your error margin in almost certainly unacceptably large.

Re:It's worse than junk (0)

Anonymous Coward | more than 3 years ago | (#31609960)

Didn't dignify the article with thinking, either, apparently

Re:It's worse than junk (1)

timeOday (582209) | more than 3 years ago | (#31610158)

"Don't dignify with a reading" = the definition of willful ignorance.

This could be as simple as watching all the text coming from an individual and getting suspicious when an "11 year old" types 90 wpm (based simply on time to respond and length of response). Would that be 100% accurate? Probably not (fire away with anecdotes about your niece), but it's entirely possible that the vast majority of kids cannot type that fast - such that it's worth collecting the data.

Re:It's worse than junk (1)

icebraining (1313345) | more than 3 years ago | (#31610326)

Besides IM programs (which are more peer to peer channels, instead of open channels), the only "chat programs" are websites, which can include Javascript to do it.

Remeber, 12 year old girls don't use IRC.

I know why! (5, Funny)

B5_geek (638928) | more than 3 years ago | (#31609580)

Is that because they are all typing one-handed? =)

Re:I know why! (1)

santax (1541065) | more than 3 years ago | (#31609622)

Where are modpoints when you need them. +1 insightful. But it does raises some concern for me. I regular type with one hand due to the other being busy handling other *output and input* tasks.

Re:I know why! (2, Funny)

cgomezr (1074699) | more than 3 years ago | (#31609764)

Yeah, we know, those where you input repeatedly until you get an output, right? You needn't have specified.

This seems like snake oil... (1)

Scrab (573004) | more than 3 years ago | (#31609606)

Even if I were to believe such a thing were possible, the fact that the Daily Mail is the source makes me roll my eyes and dismiss it... If there's a reputable source, that might be worth reading, but even then it's still a stretch to believe that all people from one culture type in the same way...

Daily Mail = Daily Fail (5, Informative)

Anonymous Coward | more than 3 years ago | (#31609608)

When will Slashdot learn?

Daily Mail is sensationalist bullshit from the UK.

Re:Daily Mail = Daily Fail (1)

vikingpower (768921) | more than 3 years ago | (#31609740)

Amen.

Re:Daily Mail = Daily Fail (0)

Anonymous Coward | more than 3 years ago | (#31609756)

When will Slashdot learn?

Daily Mail is sensationalist bullshit from the UK.

"Learn"? Slashdot is still posting stories from InfoWorld even after their stupid Windows 7 fiasco.

Re:Daily Mail = Daily Fail (1)

boneglorious (718907) | more than 3 years ago | (#31609852)

Well, mainstream reporting on science is generally sensationalistic. Even "well respected" news sources tend to say to the researcher, "How far can you possibly imagine taking this in the future?" and then publishing the answer, without any qualification, as what's being done.

Re:Daily Mail = Daily Fail (1)

Spad (470073) | more than 3 years ago | (#31610042)

That may well be the case, but that doesn't change the fact that this is from the Daily Mail, who prefer headlines of the form: "Immigrant Paedophiles in Gay Abortion Benefit Scam linked to Diana Conspiracy"

Re:Daily Mail = Daily Fail (1)

L4t3r4lu5 (1216702) | more than 3 years ago | (#31609866)

Sensationalist bullshit sells papers.

Start thinking of /. as what The Register was just before they employed Lester Haines and Andrew Orlowski, or Horizon before Jeremy Vine, and you'll see where /. is going. More "Idle" stories, more "OMG DINK OF DUR CHILDRENZ!" sensationalism, more 3 line story stubs which link back to a Digg link which links to an article from a less than reputable source.

Anybody know of an IT news website which isn't becoming more and more like the IT version of the Daily Star?

Re:Daily Mail = Daily Fail (1)

xaxa (988988) | more than 3 years ago | (#31610078)

Anybody know of an IT news website which isn't becoming more and more like the IT version of the Daily Star?

Failing that, does anyone want to start one?

Re:Daily Mail = Daily Fail (4, Interesting)

Shrike82 (1471633) | more than 3 years ago | (#31610114)

You've clearly failed to understand the important role the Daily Mail plays in British society. Every page is full of stories about how immigrants are taking all the jobs, how your tax money pays for illegal immigrants to buy pornography, how your children are at risk every time they go online to check their e-mail, how beef/bread/wine/cheese/saturated fat/unsaturated fat will WITHOUT A SHADOW OF A DOUBT kill you/reduce your risk of heart disease/prevent you from getting cancer.

The middle classes are happily distracted by the constant threat towards their very way of life, and which foods are killing them this week, that they forget about things like climate change, human rights violations around the world and the state of the global economy, which makes the government very happy indeed.

Pecking? (2, Informative)

beatsme (1472991) | more than 3 years ago | (#31609612)

It seems like the way a person was taught to type and their experience with a keyboard would determine to a large extent their speed and rhythm (and accuracy). Not to mention what they're doing besides typing on MSN or whatever. If you're drinking a cup of coffee with one hand... or switching between windows. There are way too many factors at play. Phrenology at least had the advantage of a theoretical framework. Hell, even hand-writing analysis had fewer confounding factors.

WTF?! (1)

pak9rabid (1011935) | more than 3 years ago | (#31609628)

Let me be the first to say W T F? Yeah..I'm sure this wouldn't ever produce any false positives...

I could sort off see this working. (4, Interesting)

rindeee (530084) | more than 3 years ago | (#31609632)

From a purely technological point of view, I could see this working IF you had a profile pre-built. That is to say, if you had established a baseline of a particular persons typing habits (tempo, speed, accuracy, common mistakes, pause patterns, etc.) then I think it would be relatively easy to identify/track them in the future. I realize that this isn't really what the OP is talking about, but I think this is the closest to reality as one could get. I've seen authentication tools that combine the password you've entered with the 'way' in which you enter it, doing essentially just that. Anyway, determining a persons proclivities based solely on the way they type is, well, stupid.

Re:I could sort off see this working. (1)

netsharc (195805) | more than 3 years ago | (#31609780)

There was a research once that analyze these parameters and said it can be an alternative (or additive) to password entry, they can tell by the typing speed/interval between particular letters whether the person is the same person as the previous typist or not...

Nothing new (4, Funny)

thijsh (910751) | more than 3 years ago | (#31609638)

Thats nothing, they can already measure it by the way you walk for years now... Have you ever heard of a pedometer?

Doesn't sound like pseudoscience at all (0, Redundant)

elrous0 (869638) | more than 3 years ago | (#31609646)

I'm sure that trying to tracking a person's sexual inclinations by their typing habits will prove to be every bit as useful as the science of Phrenology [wikipedia.org] .

That guy has a scary plan.... (1)

santax (1541065) | more than 3 years ago | (#31609652)

i quote: 'If children are talking to each other on Windows Live or MSN messenger, Microsoft might be able to see if there's an adult on there.' See where we going here?

Phrenology (1)

Culture20 (968837) | more than 3 years ago | (#31609682)

Tracking Pedophiles by Phrenology sounds kind of creepy. There's also a public health issue with lice and other parasites.

good luck (3, Insightful)

roman_mir (125474) | more than 3 years ago | (#31609718)

Good luck figuring anything out at all from people who are proficient enough in typing, touch-typists with years of experience who is who, who is a man, who is a woman, how old they are.

FTFA: men are 'heavy-handed and women are 'lighter' and their typing 'flows more' - that is beat by touch-typing.

But the basic problem with this is that it is a Turing test. Can a 'pedophile' impersonate someone that he/she is not? I think they should be able to, they fool kids supposedly into believing that they are also kids, of-course it is not that difficult to fool a kid, but maybe it is because it is not that hard to pretend on the Internet if you are really putting in the time to do it.

I can only guess why the 'pedophile' line is used here again, is it about the money? Is Roy Maxion trying to get a grant money from the government for this 'research'? I think so.

'If children are talking to each other on Windows Live or MSN messenger, Microsoft might be able to see if there's an adult on there.'

- MSN, the world police. What are they going to do if there is a suspicion, block access and send cops?

The department, which was formed last summer, believes the new technology could be used to prevent fraud at cash machines.

- ATMs? So from pushing 4-6 digit pin and a few menu options they are going to figure out whether it is a woman/man, what's the age etc?

I can only imagine one possibility: record information collected from the actual account holder pushing the buttons and compare the speed of typing and delays between touches and releases to try and see if it's the same person. It's not as sexy as figuring out person's age and gender though, is it?

Also how are they going to figure out how pedophiles type? Are they going to find actual pedophiles and make them sit in front of a computer and type? Did they somehow figure out that pedophiles all have the same typing abilities/habits? Do pedophiles all have the same profession and are they really that easy to figure out from the rest of the 'normal' people? Anyway, I think this 'research' is a load of bull crap and it's all about grant money.

Re:good luck (0)

Anonymous Coward | more than 3 years ago | (#31609988)

Also how are they going to figure out how pedophiles type?

That's trivial. If A claims to be a 12 year old girl in user profile, yet A types 100wpm with no errors (software assumes they are a bot or an adult), then software thinks that A is lying and hence worth investigating.

Of course you could develop countermeasures (that run as native code and trick the javascript) but use of those might be detected and make A be considered even more suspicious.

Re:good luck (1)

roman_mir (125474) | more than 3 years ago | (#31610030)

THAT is the whole point of not being an idiot, or do you assume that all pedophiles are also stupid? They fit in, they pretend being something they are not and many of them have to be good at it, so they would not behave like adults would and that includes the typing abilities and patterns, the vocabulary, the cultural references, everything.

How do you miss that from my original comment?

Re:good luck (0)

Anonymous Coward | more than 3 years ago | (#31610074)

THAT is the whole point of not being an idiot, or do you assume that all pedophiles are also stupid?

Seeing as they're invariable catholic priests who believe in mythical deities and magical people, I'd say it's a good starting point.

Re:good luck (1)

roman_mir (125474) | more than 3 years ago | (#31610132)

However catholic priests don't need to do any typing to get what they are looking for, this software ain't for those, is it?

Looks like this is simply an attempt at a grant for a useless research that will accomplish nothing. Those who are on line and are hunting kids, those have to be able to fit in, to pretend to be something they are not. Figuring them out by their typing skills among tens of thousands of other participants of chat boards? Give me a break, it's bull shit.

so ummmm... (0)

Anonymous Coward | more than 3 years ago | (#31609762)

So the root of this plan is to install a keystroke logger on all computers. They should just use the webcam. ARF

Another interview on this story (1)

Troy (3118) | more than 3 years ago | (#31609772)

On the Media (a fantastic program if you haven't heard of it yet) covered this last month. Interview and transcript are here: http://www.onthemedia.org/transcripts/2010/02/26/03 [onthemedia.org]

Correlation, implication, causation etc. (5, Interesting)

Rah'Dick (976472) | more than 3 years ago | (#31609790)

They say they can identify a single person by a typing profile that they've previously generated, but you certainly cannot deduct from a typing profile that any given person is a pedophile! I agree with the author's comment.

[rant]On a side note, TFA has pictures of a murderer and a 17 y/o girl next to each other. I refuse to call someone a "pedophile" who is attracted to 17-year-old girls, because becoming an adult is not a matter of some age number increasing by one digit. 17-year-olds are certainly not KIDS! I'm sick of the misuse of the word "pedophile". What they actually mean is "ephebophile", but since that one is - to some degree - accepted by society, they cannot polarize people enough to enact more stupid laws. [/rant]

Re:Correlation, implication, causation etc. (1)

boneglorious (718907) | more than 3 years ago | (#31609894)

Well, remember, they're not tracking a feature, "Is a pedophile," they would be tracking a feature "is an adult" and a feature "is in a child's chat typing suggestive messages".

Re:Correlation, implication, causation etc. (2, Insightful)

mdwh2 (535323) | more than 3 years ago | (#31610138)

I fully agree with your rant - and furthermore, this article is in the UK, where 17 is over the age of consent.

It's rather ridiculous - that story has been all over the media, and referenced in stories to do with child abuse, such as the recent demands to force Facebook to have a big red panic button. Not just the tabloids either - places like the BBC have been reporting the murder story, as if she was underage. Yet not one of these places has even noted that the age of consent is actually 16.

It wouldn't have mattered if she was 17, 18, or 50. The problem here was not how old he or she was, but instead the rather more serious point that he murdered her. (I think it's telling that a simple murder is not considered bad - no, we've got to report on someone being 17, even if that's not underage...)

Re:Correlation, implication, causation etc. (1)

xaxa (988988) | more than 3 years ago | (#31610302)

The age of consent in the UK is 16 anyway, also the age when you can leave home.

Don't expect the Daily Mail to bother with the facts though, when they can do sensationalism instead.

Great (1)

vadim_t (324782) | more than 3 years ago | (#31609844)

Of course, there's no way children and adults could ever legitimately interact anywhere online. They never happen to play the same games, post to the same forums, and parents never talk to their children online.

Re:Great (1)

boneglorious (718907) | more than 3 years ago | (#31609916)

Yes, but hopefully those parents don't say things like, "What are you wearing?" (followed, of course, by, "Go take that off immediately, young lady, you know you aren't allowed to dress like that while you're living under my roof!" :D )

Re:Great (1)

boneglorious (718907) | more than 3 years ago | (#31609938)

The point being that it probably isn't too difficult to tell the difference between a person trying to reach an appropriately sexual result and one who's interacting permissably with children.

95 percent accuracy isn't a lot (1)

91degrees (207121) | more than 3 years ago | (#31609874)

For every 1000 people, 50 will be misidentified. Unless there are a lot of paedophiles this will make it essentially useless. Although really this is the Daily Hate tacking the story onto one of their favourite scares.

ffs (4, Insightful)

Therilith (1306561) | more than 3 years ago | (#31609918)

Pedophile != child molester

better chance (1)

roman_mir (125474) | more than 3 years ago | (#31609926)

I think the research would get better results if they concentrated on using the web-cams to see who is in front of the computer. Shouldn't be too hard, they do this at US schools already, just supply everyone with a 'free' laptop and disallow anyone from using their own equipment on the web and make sure that the camera on the laptop can be used any time at all to view the laptop user.

Then just look for clues [youtube.com] to figure out who is a pedophile. Job done.

Can I have the grant money to do this now?

don't forget the beards (1)

roman_mir (125474) | more than 3 years ago | (#31609954)

Combine the pedophile beards [youtube.com] with the rapist glasses, the kind of look that women love.

Keyboard pattern and identity (1)

woboyle (1044168) | more than 3 years ago | (#31610002)

I did some research into this back in the early 80's, using a person's typing patterns to identify them and worked up the algorithms necessary to implement a 2-factor authentication system, a passphrase in combination with an analysis of the typing pattern as the user inputs the phrase. The main obstacle (human factors) was the training necessary to get a good pattern to compare against. Computers at the time were a lot slower and less capable than they are now. This can work, but it has a LOT of gotchas, such as what happens when you are injured in the extremities and your pattern changes significantly?

Re:Keyboard pattern and identity (0)

Anonymous Coward | more than 3 years ago | (#31610112)

The article mentions that it can identify someone with a hand injury, so I assume they would just throw out that result. The number of pedos with hand injuries is probably pretty low.

So... where is the demo? (1)

clone53421 (1310749) | more than 3 years ago | (#31610014)

Anybody have a link so I can try it out? I wanna see how it identifies me.

What... no! Of course I’m not a pedophile...

Spoiler, I read TFA... (2, Informative)

clone53421 (1310749) | more than 3 years ago | (#31610100)

Apparently they had to hook up sensors to the typists’ fingers. Now we just need to have all the paedophiles register so that we can make sure they’re wearing their typing gloves when they go online.

(Why they couldn’t just use a regular keylogger is beyond me, but the fact that they didn’t implies that for some reason a simple keylogger wouldn’t work. It would certainly have been easier than wiring up the typists and developing electronics to monitor the sensors.)

Remember graphology? (3, Informative)

jazcap (1125477) | more than 3 years ago | (#31610108)

For a time in the '70s, World's Fairs, science expos, etc. might feature handwriting analysis by computer. A large, impressive, punchcard-based machine would read your signature or other writing sample and produce a 'description' of your personality.

Graphology has at times significant support, and its use has been explored by criminologists. Nowadays, it's generally considered to be a crock. One of the areas where graphology was earliest discredited was in its (in)ability to tell the gender of a writer.

The Daily Wail article makes similar claims to those made for graphology, including the ability to determine gender, and proposes some of the same uses for the new technique. I suspect this will wind up on the crock-shelf next to graphology.

Great....just started to file a patent application (1)

Tanuki64 (989726) | more than 3 years ago | (#31610180)

Keyboard proxy. A device which is plugged between keyboard an computer and which accepts key strokes and sends them in a freely selectable pattern to the computer. The base package contains: 'young white woman', 'trustworthy, well situated man', and 'perfect lover'. Additional packages like 'honest politician' or 'horny milf' can be ordered separately.

So will this be as speechrecognition? (1)

ZeroExistenZ (721849) | more than 3 years ago | (#31610256)

Depending on my mood alone, I have a different typing style. At work I have a different chair which forces a more upright posture, which changes my typing.. also the type of thing I would be typing: if I'm inspired my typing is very fast and in a continious stream, in a chat it's often less quick.

I doubt there is anyone who has a continious typing habit. It's just a curiosity and technical exercise, the "pedo"-tag helps them justify their work

Rythym? (1)

Aardpig (622459) | more than 3 years ago | (#31610280)

What, like everyone has a unique 'wanking fingerprint'?

Hmm... (1)

scuzzpuck (874594) | more than 3 years ago | (#31610288)

Surely they just need to install sensors that are sensitive to the smell of hammers?

Very misleading article (1)

GrumpySteen (1250194) | more than 3 years ago | (#31610290)

The research mentioned in the article is about verifying identity by comparing a person's keystroke timing to a previously acquired of keystroke timing. Nothing the professor's research makes any claim of being able to determine age, gender, culture or anything else from an unknown person's typing.

At best, you could compare a person's typing to a database of typing patterns and see if you find a match. To identify pervs this way would require that you track down all the pervs, give them typing tests, ensure that they type like they normally do, store the typing profiles in a database, ensure that all programs record and send the timing of typists keystrokes, ensure that the pervs all type the same way they did on the typing test and compare the resulting keystroke timing profile.

But hey... how could any of that possibly fail to work?

Also, the professor's list of published papers can be found here. There is some interesting stuff there
http://www.cs.cmu.edu/~maxion/pubs/list.html [cmu.edu]

Circumvention (0)

Anonymous Coward | more than 3 years ago | (#31610328)

This could be easily circumvented by a very simple program that uses the then-not-so-secret[1] typing profiles to first buffer my input and then emulate the keystrokes in the desired fashion while passing it to any chat client or whatever.

Epic fail.

[1] As an alternative I can trick some relatives or friends into handing out their personal typing profile to me.

(Sorry for my bad english)

Load More Comments
Slashdot Login

Need an Account?

Forgot your password?