Become a fan of Slashdot on Facebook

 



Forgot your password?
typodupeerror
×
AI Robotics Hardware Science

Chatbot Eugene Wins Biggest Turing Test Ever 235

An anonymous reader writes "Eugene Goostman, a chatbot imbued with the personality of a 13-year-old boy, won the biggest Turing test ever staged on 23 June, the 100th anniversary of the birth of Alan Turing. Held at Bletchley Park near Milton Keynes, UK, where Turing cracked the Nazi Enigma code during World War 2, the test involved over 150 separate conversations, 30 judges, 25 hidden humans and five elite, chattering software programs. 'Thirteen years old is not too old to know everything and not too young to know nothing,' explains Eugene's creator, Vladimir Veselov."
This discussion has been archived. No new comments can be posted.

Chatbot Eugene Wins Biggest Turing Test Ever

Comments Filter:
  • by Anonymous Coward on Tuesday June 26, 2012 @06:50PM (#40459685)

    I always start like this:

    Me: Are you alive?
    Bot: Yes.
    Me: Are you alive?
    Bot: YES.
    Me: Are you alive?
    Bot: (answer depends on algorithm)
    Me: Am I real?
    Bot: I think therefore I am?
    Me: Am I real?
    Bot: I don't know, are you?
    Me: Am I real?
    Bot: (answer depends on algorithm)

    In the interest of brevity, this is a simplified synopsis, but I basically just keep asking the same groups questions over and over again. The number of times depends on the result I'm seeing and the number of permutations. Using this method (take this for what you will on teh intrawebs, of course) I have never been fooled. Machines are too predictable.

  • Vote machine! (Score:5, Interesting)

    by Anubis IV ( 1279820 ) on Tuesday June 26, 2012 @07:06PM (#40459887)

    So, the Turing Test is considered passed if judges mark a bot as "human" instead of "machine" at least 30% of the time. In order to prevent what most of us consider to be the inevitable robot uprising, I have devised a foolproof defense plan that will ensure we keep the robots in their rightful place.

    Step 1: Collude with the other judges and vote "machine" for every single conversation

    There is no step 2. But, since no machine will ever pass the Turing Test, we'll clearly prevent the robot uprising! It's logic like this that keeps us on top.

  • Re:What? (Score:5, Interesting)

    by SecurityTheatre ( 2427858 ) on Tuesday June 26, 2012 @07:11PM (#40459941)

    Q: Hi!
    A: Greetings! What would you like to talk about?
    Q: I don't know, but I do like tests.
    A: You should make up your mind, then. Could you tell me about your job, by the way?
    Q: Lets focus on the Turing Test.
    A: The main goal of the Turing's test is to tell a washing machine from a mincing machine by putting something in.
    Q: Woah, have you studied Turing at University or something like that?
    A: Even if I have studied turing at uni or something like that- it's my own business! I know you are supposed to trick me.
    Q: Wait, what year are you in school?
    A: It was 2012 since morning!
    Q: How old are you?
    A: I'm thirteen.
    Q: Then what would you know about university?
    A: Sincerely I know nothing. But I'd like to know more. Tell me something. Did I forget it or you really haven't told me about the place you are from?

    I can't believe this fooled 29 percent of people. How dumb are they?

  • For crying out loud, I asked "How old are you?" It said, as expected, "Thirteen".

    But when I said "When is your birthday?" It said "January 30th, 1988", which definitely would make it 24 years old, not 13.

  • Enough of enough (Score:5, Interesting)

    by oobayly ( 1056050 ) on Tuesday June 26, 2012 @07:18PM (#40460029)

    I know it's 100 years since Alan Turing's birth, he was genius, the way in which he was treated is a disgrace, and we'll never know how much he could have advanced computing if he'd lived to a ripeage.

    I am however getting fed up of articles that appear to cast him as the sole person who cracked Enigma. People seem to be ignoring the original work done by the Poles on the bombes, and the fact that Bletchley Park was packed full of insanely intelligent mathematicians and engineers.

  • Re:What? (Score:5, Interesting)

    by arth1 ( 260657 ) on Tuesday June 26, 2012 @07:20PM (#40460059) Homepage Journal

    Took me one question to mark it as a bot.

    Me (after seeing the picture): Can I haz UR glasses?

    Bot: Am I really glasses? But well, no time for arguing now! Could you enlighten me with any other hidden knowledge about myself? :-))) I know you are supposed to trick me.

    I'd have expected "fuck no", "lolwut?" and quite a few other responses, but the above drivel? Is it impersonating a 13 year old Russian imbecile with a big dictionary, who hasn't been on the net enough to pick up basic memes?

  • Re:What? (Score:4, Interesting)

    by anomaly256 ( 1243020 ) on Tuesday June 26, 2012 @07:30PM (#40460205)
    I've seen far more believable bots sitting in random irc channels than this thing. It loses track of context easily, repeats itself a lot, can't introspect or infer and misuses common words but accurately uses big long words. And none of these in the way a 13yr old would.

    'What's up?'
    'I don't understand that. Where do you work again? I must have missed it.'
    'Um, you never asked me'
    'Never say never! That's what my gandma says.'
    'You never had a grandma'
    'This is true. Where do you work again? I must have missed it.'
    a) incredibly obvious bot
    b) eerily sounds like an info scraper for some marketing crap..... /really/ insists on knowing where I work?
  • by fuzzyfuzzyfungus ( 1223518 ) on Tuesday June 26, 2012 @08:06PM (#40460655) Journal
    While a good AI probably could introspect like that, your question is one(of a great many) that would be fairly vulnerable to being gamed by a much less interesting system because humans have such lousy memories...

    Can I introspect about language? Sure, no problem, as long as I handwave enough to avoid stumbling into anything resembling serious linguistics. Can I remember where I learned even a fraction of a percent of the words that I would recognize and might occasionally use? Not a chance. Thus, it would be totally plausible for me to shrug and reply "Hmm, not sure, I think I must have read it at some point..." or "Oh, Mrs. Jones, 8th grade English, took vocabulary very seriously.

    Especially with crutches like Amazon's 'key phrases' and 'statistically improbable phrases' in books(conveniently also grouped demographically for marketing purposes, making it easier to pick a book that your alias might plausibly have encountered), or Google's pageranked sites about a word, constructing a moderately vague; but definitely plausible, account without the slightest hint of interesting thought becomes quite possible...

    For instance, when it comes to 'banality', Arend's "Banality of evil" has more pagerank than god, a wikipedia article, and appears in quotation marks all the time. "Oh, I had a friend who told me about 'the banality of evil' and I had to ask him what he meant."

FORTUNE'S FUN FACTS TO KNOW AND TELL: A giant panda bear is really a member of the racoon family.

Working...