3 ms·
The whole thing sounded iffy to me from the first second I heard about it, despite the fact that most of the tech media was going ga-ga over it. I mean, the AI
by higherpurpose 12y ago
The whole thing sounded iffy to me from the first second I heard about it, despite the fact that most of the tech media was going ga-ga over it. I mean, the AI being only a teenage Hungarian boy (meaning the judges should've expected it to be "less intelligent than a normal native adult")...and only fooling 1/3 of the judges? That sounds like a whole lot of cheating to declare that it "passed" the Turing test.
- oxalo 12y agoYeah after reading Ray's conversation with the bot, I was surprised that 30% of the judges were fooled. I guess maybe they were also 13 years old and had English as a second language.
- ma2rten 12y agoI will explain this mystery to you: they simply did not ask any probing questions like that. This is also called the ELIZA effect http://en.wikipedia.org/wiki/ELIZA_effect http://en.wikipedia.org/wiki/ELIZA_effect
- bentcorner 12y agoYou are now chatting with babybot You:> Hi, how are you? Babybot:> asdfasdfu3898bzasgasdlasdl3897a9sdf79as87fd
- orbifold 12y agoPretty much any question that requires actual natural language understanding would presumably instantly reveal a bot today, like 'Is white darker than black?'.
- Houshalter 12y agoIt's not that simple. NLP has gotten good enough that there actually are systems which can answer questions like that. IBM's Watson can quickly sort through a massive number of documents for an answer. There are projects that have built massive databases of common sense knowledge like Cyc, which could possibly answer that question. And then large databases of human responses like cleverbot which probably already has been asked that at least once and can find it in it's database.