Forgot your password?
typodupeerror
Software Privacy

Why Your Online Impersonation of a 16-year Old Girl Won't Last Long 137

Posted by timothy
from the because-soon-enough-you'll-be-17-anyhow dept.
An anonymous reader writes "Can computers pick up your age and gender from your tweets? If you want to give it a try, here's your chance: 'To develop your software for age and gender identification, we provide you with a training data set that consists of blog posts, Twitter tweets, social media texts, as well as hotel reviews.' Well, at least my paid Amazon reviews are safe for the time being..."
This discussion has been archived. No new comments can be posted.

Why Your Online Impersonation of a 16-year Old Girl Won't Last Long

Comments Filter:
  • assuming too much (Score:5, Interesting)

    by roman_mir (125474) on Thursday February 20, 2014 @07:31PM (#46300103) Homepage Journal

    I am sure you can pick up on general mood of a person, I am sure you can pick up various clues, but if somebody is set on hiding themselves by providing false information, I don't think you'll be able to identify them without what NSA calls 'meta data', as in pattern of your behaviour. I don't think it will be possible to do a better job than guessing at about 50/50 chance ratio about a person's age and gender if the person in question is actively trying to portray something he or she is not in a single conversation. From a pattern of behaviour? Yes. From a single conversation? ... this assumes too much about people. It assumes that 16 y.o. girls are also not pretending to be something they are not as well...

    • by rasmusbr (2186518)

      Men v.s. women ought to be fairly simple as long as we're dealing with untrained writers. Women use social words like pronouns and verbs about people more often than men. Women are also, arguably, slightly better writers than men on average because they make more of an effort in primary school.

      • by AnttiV (1805624)

        This is generally true (probably, I haven't done any scientific testing) and because of that any analysts and/or analytic program assume as much - which is why almost 95% of the time those instances define me as a female based on social behaviour in games/chats and texts that I write (stories, etc).

        So while it is fairly simple to do it that way, it is also assuming too much as people are rarely, if ever, so one-sided that any clinical analysis could determine their age or sex short of looking into their pan

        • by Nutria (679911)

          which is why almost 95% of the time those instances define me as a female

          Well, 5% of men like what women like, so that shouldn't be terribly surprising.

      • by Sique (173459)

        Women are also, arguably, slightly better writers than men on average because they make more of an effort in primary school.

        ... which just means that an slightly above average man can impersonate an average woman.

        • by rpstrong (1659205)

          ... which just means that an slightly above average man...

          That should be a, not an. You must be a guy.

      • by geekoid (135745)

        What a pile of crap.

      • by mpe (36238)
        Men v.s. women ought to be fairly simple as long as we're dealing with untrained writers. Women use social words like pronouns and verbs about people more often than men. Women are also, arguably, slightly better writers than men on average because they make more of an effort in primary school.

        IIRC the "Turing Test" was originally posed as if it were possible to tell the difference between a man and a woman. Rather than between a human and a computer.
        The other issue is that someone attempting to decieve m
    • Re:assuming too much (Score:4, Interesting)

      by canadiannomad (1745008) on Friday February 21, 2014 @04:08AM (#46302269) Homepage

      Even if someone got a program that was "pretty good" at figuring out age and gender, I strongly doubt they could make it accurate enough not to be a huge civil liberties problem.
      IF someone were to make a program that was 99% accurate, that would still be some 3million 16 year old girls in the US getting their houses raided in case they are a perv. (ok they aren't likely to raid on evidence like that alone, but they might use it as probable cause to 'dig deeper' into who she is... and if she caused the false flag then she is likely an outlier who the gov't wants to know more about anyway, for retraining... I'll go don my tinfoil hat now.)

      • by TheCarp (96830)

        Well, exactly.... its a base rate fallacy.

        The false positive vs flase negatives are two different error conditions. A 95% accuracy is great, if you have only 4 suspects, and you are pretty sure one of them has to be the one, and a 95% accurate test separates out one of them.... awesome.

        However, unless there are many more people pretending to be 16 year old girls than 16 year old girls, which seems highly unlikely.

        That said, judging age can be hard. There was a small group of people that I used to IRC with i

    • by flyneye (84093)

      Yeah, but who cares what a 16 yr. old girl has to say about anything?

  • by Anonymous Coward on Thursday February 20, 2014 @07:34PM (#46300137)

    here's your chance to write one.

    • Slashvertisement! Slashvertisement! Slashvertisement! Slashvertisement! Slashvertisement! Slashvertisement! Slashvertisement! Slashvertisement!

      Shady competition seeking to pay pittance for million+ Euro development, read all about it!

      Slashvertisement! Slashvertisement! Slashvertisement! Slashvertisement! Slashvertisement! Slashvertisement! Slashvertisement! Slashvertisement!
    • by pitchpipe (708843)
      On the classic slashdot site the summary made sense!

      Nah, I'm kidding: even there it was gibberish.

  • by EmagGeek (574360) <gterich@aoMONETl.com minus painter> on Thursday February 20, 2014 @07:35PM (#46300141) Journal

    I wish I could shell out 300 euro to have my next million-euro commercial product crowdsourced

  • Why do we have to help them train the replacement twitter scanners after the EU is trying to cut off their direct pipeline?

  • by Anonymous Coward

    How are they going to catch pedophiles now?

    • by Garridan (597129)
      Hire and train real 13 year-old girls to seduce 'predators' online and lure them into an IRL meeting. Duh.
      • by allo (1728082)

        I guess its the same problem. A girl hired for such a trap WILL behave other than a normal girl. maybe its not even recognized by the software then.

    • Must have been...15 years ago or so, there was an SNL skit involving chat rooms in which a naked and hairy Chris Elliot pretended to be "Clair the filthy slut," a 16 year old cheerleader. Whenever I see anyone online claiming to be a young girl, I immediately think of a naked and hairy Chris Elliot.

  • by mbone (558574) on Thursday February 20, 2014 @07:41PM (#46300187)

    For those puzzled by the description here, this is a software contest, with a 300 Euro prize :

    An award of 300 euros for the best performing approach to author profiling (age and gender identification) is sponsored by Atribus (Corex).

    • Re: (Score:2, Funny)

      by Anonymous Coward

      What they forget to mention is that if your program also identifies ethnicity, you will be fined for racism.

  • by Anonymous Coward

    >Why Your Online Impersonation of a 16-year Old Girl Won't Last Long

    This is one of those posts where I have to read the summary, isn't it.

  • So you like to chat with 16 year old girls on line?

  • by Dan Askme (2895283) on Thursday February 20, 2014 @07:52PM (#46300265)

    I'am sick of descriptions and subjects written in gibberish. Reminds me of spam emails.

    Do you have a big cock, how big can your cock get?
    Win £xxx pounds if your cock can be bigger than their cock.
    Is your cock hidden? can people see your cock? CLICK HERE NOW!

    How the fuck did this get authed? Not only is it the worst subject and description i've seen in a while, its an insult to the readers here. Seriously we are here for a reason, we dont have an attention span of a fish.
    Oh yeah thats right, its a Slashadvert supplied by Corex.

    No offence, but if this shit carries on, i may as well start reading my spam emails. Target your audience, not Facebook muppets ffs.

    • I'am sick of descriptions and subjects written in gibberish. Reminds me of spam emails.

      Do you have a big cock, how big can your cock get?
      Win £xxx pounds if your cock can be bigger than their cock.
      Is your cock hidden? can people see your cock? CLICK HERE NOW!

      How the fuck did this get authed? Not only is it the worst subject and description I've seen in a while, its an insult to the readers here. Seriously we are here for a reason, we don't have an attention span of a fish.
      Oh yeah thats right, its a Slashadvert supplied by Corex.

      No offense, but if this shit carries on, i may as well start reading my spam emails. Target your audience, not Facebook muppets ffs.

      For the record, my rooster is the biggest in the entire state. However, when I post pictures of it people get deeply offended. They sware I stroked by rooster so it stands taller and more robust. Hey, how else are you to get fresh eggs from your hens but by having the most spirited rooster. So Thats what I always send those spam emails. Never won any prizes though. As for the topic, Twitter will reveal all you really need to act like a teen if you follow enough of them.

  • Researcher A: This... This c-can't be right!

    Researcher B: What's not right?

    Researcher A: Apparently, 4chan really is populated by little girls.

    Researcher C: *posting anonymously on /a/* See guys, I told you -- wearing a skirt makes it work. Also, Check'Em.

  • According to the research, 83% of slashdot posters are 16yr old girls. Go figure.

    • by Anonymous Coward

      According to the research, 83% of slashdot posters are 16yr old girls. Go figure.

      Well that explains Slashdot beta doesn't it.

  • by knorthern knight (513660) on Thursday February 20, 2014 @08:27PM (#46300545)

    The "dialectizer" http://www.rinkworks.com/diale... [rinkworks.com] "translates" English to Redneck, Jive, Cockney, Elmer Fudd, Swedish Chef, Moron, Pig Latin, or Hacker. And there's an English to Ebonics translator at http://joel.net/EBONICS/Transl... [joel.net] so it won't be that difficult to get a translator that outputs 16-year-old-girl talk.

    • Your comment run through the "Jive" filter at the link you provided:

      Your Text, Dialectized (jive)

      De "dialectizer" http://www/ [www] so cut me some slack, Jack.rinkwo'ks.com/diale... [rinkwo'ks.com] "translates" English t'Redneck, JIBE, Cockney, Elma' Fudd, Swedish Chef, Mo'on, Pig Latin, o' Hacker. Ah be baaad... And dere's an English t'Ebonics translato' at http://massa/ [massa] l.net/EBONICS/Transl... [Massa' l.net] so's it won't be dat difficult t'get some translato' dat outputs 16-year-old-goat rap.
      --

      I'm not repeati

  • Seems like it will be based too much on stereotypes. When I was 16, I typed in proper case (I wanted most of all to be a writer, and I used every opportunity to improve my grammar and typing), used big words in context, and did not easily use emoticons.

    I suppose over time they could develop something keener, but if you're going to base it on the mean of all people in a certain age bracket, there will be enough exceptions to render the most useful applications of the software irrelevant.
  • I could see a bunch of ways to make tons of money from this, starting with selling it to FaceBook for $19B.

    Why would I publish it for 300 Euro again? I know they *claim* it's not published, but if they didn't sign an NDA, you're not going to get a patent out of it outside the U.S., and you're not going to have any protection against them just using your algorithms.

    This is a really silly contest.

    • Three hundred euro? The contest sparks my interest, but 300 is about what it would take to get me to fill out the entry form. To develop an effective NEW algorithm, code it, and test it in HOPES of winning the prize? Maybe for 300,000, maybe. 3,000,000 would be more like it.

      I've developed exactly two truly innovative products. One I sold over $1 million worth, the other still provides $3,000 / month in net income . Why would I, or anyone skilled and innovative, touch this for 300 euro?

  • Didn't Gender Genie already attempt to do this? https://www.google.com/search?... [google.com]
    (I say "attempt" because I found that even in cases where I wasn't trying to fool it, it would often come up with the wrong gender.)
  • by godrik (1287354) on Thursday February 20, 2014 @10:37PM (#46301307)

    A friend of mine was working on sentiment analysis. They studied the content of yahoo answer and it was quite interesting all the correlations that you can make. The study is of course not enough to provide a direct identification, but it shows how many parameters you need to keep in mind when building a "virtual identity".
    http://www.cse.ohio-state.edu/... [ohio-state.edu]

  • by MacTO (1161105) on Thursday February 20, 2014 @10:54PM (#46301367)

    Now both people and computers will call me a girl.

  • I managed to get my impersonation of a 16-year-old girl into the training set.

  • by dietdew7 (1171613) on Friday February 21, 2014 @01:30AM (#46301857)
    1. You can assume that someone is female if they act like a man without reason or accountability. 2. Anyone that shows interest in you sexually is an undercover police officer.
  • What kind of person is willing to spend 300 euros for detecting 16 year olds online, mmmm?
  • Computers are very bad at judging these things. It isn't their strength.

  • You are going to have to pay me more than 300 euro for that. ~400 should do it.
  • Just read the competition, and thought I should point this out:

    participants will be asked to classify the author of a set of tweets as journalist, politician, activist, professional, client, company, authority or citizen, since the fact of belonging to a certain category could determine the importance of the user's opinions.

    If you submit a solution, know that it will be used to classify journalist and activist communications.
  • What's my motovation to hide as a 16 year old girl?
    • What's my motovation to hide as a 16 year old girl?

      To troll some nsa perverts into using their laptop camera exploits into think they are getting are going to spy/perv on a 16 year old female and instead get a naked neckbearded broney wearing a pantomime hoarse head while pulling a goatsx?

  • Look at the original proposal by Turing. I imagine Turing was very conscious about gender issues since after winning wwii for the old boy network they decided he was a felon. It seems to me the five eyes gives us Turing test success and...wait for it ... means that is not really what we meant by human intelligence.

  • Togged to the bricks, and all they offer me is a trip for biscuits. Just wanted a ring-a-ding ding and some Bruno slips me a Micky Finn in a clip joint. Next thing I know, I'm waking up in a flop with some Lunger wearing iron who suddenly runs to the window and starts drilling beans at a tin can full of Joes. So I make tracks before the coppers turn up with a meat wagon and someone ends leaving in a Chicago overcoat.

Uncompensated overtime? Just Say No.

Working...