During a Turing test a human judge will chat with something hidden away, usually in a different room. The judge is not supposed to see, or hear the entity, they must use type written communication only to ask questions and then decide if they are talking with a human or a machine.
The purpose of the ‘hidden human’ is to be themselves when they answer questions put by the human judge. The purpose of the machine is to provide satisfactory answers to the human judge’s questions.
Sometimes human judges think they are talking with a machine, but in fact it was another human, this is because the human gave odd replies which the human judge thought were not human-like. Other times human judges think they have been talking with a human when in fact it was a machine – in this case the human judge felt the answers were satisfactory.
There are times when the human judge really can’t decide whether they are talking to a human or a machine, because the answers are not what the human judge might have expected.
If you would like to read more, please see my Academia.edu page, you can download and read my PhD thesis on ‘Deception-detection and machine intelligence in practical Turing tests’:
If you would like to take part in a future Turing test experiment do keep in touch via my blog here, I recently blogged that the Turing test is not just ‘theatre’: http://humashah.blogspot.co.uk/
Hope this answer helps to explain what heppens in a Turing test.
Comments