How online 'chatbots' are already tricking you
June 10, 2014  03:27

Intelligent machines that can pass for humans have long been dreamed of, but as Chris Baraniuk argues, they're already among us.  

 

Sometimes it's the promise of sex that fools you. Sometimes it's because they seem wise, friendly or just funny. The bots don't really care how they trick you '" their only objective is to make you think they're human. In fact, if you use social media or spend any time online, it's quite possible you've already been a victim.

 

This week, a controversial claim was made that a 'chatbot' passed the Turing test at an event at the Royal Society in London. During a series of text-based conversations, a computer program named Eugene Goostman persuaded judges it was a 13-year-old Ukrainian boy, thus passing a benchmark for artificial intelligence proposed years ago by the computer scientist Alan Turing.

 

So does this announcement mark the era of human-like AI, as has been claimed? Not really. Turing's test stopped being important for AI research years ago, and many scientists see the contests as flawed because they can be won with trickery '" such as pretending to be a non-native English speaker.

 

Read more on BBC

« Back to LIVE

TOP STORIES