How many movies, TV shows or video games have you seen with a computer that could think, talk, and have a conversation with other characters? HAL-9000 from 2001: A Space Odyssey. The Cylons in Battlestar Galactica. Mr Smith in The Sarah Jane Adventures. Cortana in Halo. Johnny Five in Short Circuit—the list is endless.
We now have software that can follow spoken instructions, or hold basic conversations to finish a task. You might even have one in your phone, such as Siri or Google Now. But can they really think?
In 1950, Alan Turing, a pioneer of computing, came up with a twist on a party game called “The Imitation Game” that begins to answer that question.
Imagine a conversation in a chat room between three different people. In real life, these people are in different rooms, so can’t see each other as they chat. And in fact, one of the people isn’t a person at all, but a computer, designed to talk and behave like a human would in a chat room.
If the third (human) can’t tell which of the other two is the computer and which is the human, then the computer has ‘passed’ the Turing test.
But some scientists disagree on whether, if a computer passed the Turing test, that would mean the computer was really ‘intelligent,’ or if it had a ‘mind’ of its own.
In 2014, a Russian chatterbot program called Eugene Goostman apparently ‘passed’ the Turing test in a competition organised by the University of Reading. ‘Eugene’ pretended to be a Ukrainian teenager who was learning English, rather than an adult with English as their first language. He also avoided answering most of his questions by changing the subject. In the end only one third of the competition judges were fooled by Eugene
Many people don’t think it really did ‘pass’—but he still made headlines around the world.
Clever critterness: but not quite clever enough
More AI stuff →
A bit of cs4fn 