- Seamus McGrenery
- Dublin
- Ireland
This conversation is closed. Start a new conversation
or join one »
Could the Turing test, as originally posed, be impossible for a machine?
When Alan Turing sat down to devise a test of whether a machine could be described as intelligent he choose to describe a test based on the imitation game.
‘The new form of the problem can be described in terms of a game which we call the 'imitation game'. It is played with three people, a man (A), a woman (B), and an interrogator (C) who may be of either sex. The interrogator stays in a room apart from the other two. The object of the game is to determine which of the other two is the man and which is the woman.’
If, as I believe, intelligence can be described as knowledge applied for a purpose then this test poses an impossible challenge for any machine.
The real purpose of the imitation game is to test the interrogators ability to identify which of A or B is a potential rival and which a potential mate.
While a machine could be programmed to take one of the roles in the game, it has no real stake in the ‘mate or rival’ question.
Can a machine can ever have an independent internal purpose?
Closing Statement from Seamus McGrenery
Thank you to everyone who took part in the conversation.
The reason I asked the question was that I was struck by how, in refining the description of a test for machine intelligence, Alan Turing had used an example of a binary choice that is meaningless for machines - that of mate or rival.
Artificial is a word we humans often use to describe the things we make. The latin root of the work is skill. We are of course impressed by the skill of our species in creating all manner of machines.
Humans are animals and, as far as I can see, what we do is ultimately motivated by the survival of our selves, or families, our species.
In evolutionary terms we are living in a period of
extraordinary rapid change. In less than one hundred thousand years our species has become the dominant one on the planet. But the last hundred years has seen our numbers double, then double again. We have become very adept at making tools to promote our peaches. As animals we must be doing something right.
Maybe we should see all of our tools, including computers, first and foremost as things which are helping us to thrive.
We have had less than a century to get used to computers. In gaining an understanding of them it was natural to think terms like 'electronic brain' or electronic mind.
Yet if the computer is really an artificial replication of our brain, then why are we so poor at math. Maybe that is a different discussion.
Showing single comment thread. View the full conversation.
Showing single comment thread. View the full conversation.
Barry Palmer 50+
Is it possible? Yes, of course, if someone puts it into the program.
Is it desirable? Absolutely NO. This must be avoided.
Intelligent machines must be limited to answering questions. Giving them any kind of internal purpose could result in loss of control. The moment such a machine becomes smarter than we are, it becomes completely unpredictable.
By the way, I would not consider the 'imitation game' as a new form of Turing's test. As you point out, the imitation game requires an internal purpose, whereas Turing's test did not. This is a very significant qualitative difference.
Seamus McGrenery
As his biographer put it 'Turing wants to argue that the successful imitation of intelligence is intelligence'.
Turing's view is that any feature of the brain relevant to thinking or intelligence can be decried as a 'discrete state machine'.
He was certainly advocating the idea that intelligent machines could be built.
However maybe something about the quality of his own intelligence led him to write the aspect of the test requiring an internal purpose.