Robots with advanced AI features are being developed to deal with a multitude of functions.

We have long been used to the idea that human beings represent the pinnacle on the evolutionary scale, and we have drawn the conclusion, at least unconsciously, that nothing could possibly happen to topple us from this dizzying height.

And yet it is enough to look around us to find that in many aspects, there are other creatures who exceed our physical capacities in one or more ways: eyesight is much better in birds; the sense of smell is a thousand times more acute in dogs, bats can manoeuvre around objects and home in onto a mosquito in complete darkness.

But we were hoping that at least in one faculty, intelligence, we can still claim to be supreme. Yet even this concept is now being subject to revision and it is quite possible that in the not-so-distant future we will be challenged by a new range of non-human ‘individuals’ with higher intelligence that may take over the world. Such a development will certainly raise questions which may need answers.

We are all aware of the enormous progress that has taken place in artificial intelligence (AI) in just the last few years. Aspects of mental life which we always associated with intelligence have now been taken over by machines that can manipulate figures faster, have a far superior memory that cannot be matched by the best brain, have the capacity to navigate a car or an airplane as well as a human being.

Most worrying of all, perhaps, is the newly-acquired ability of learning from their own mistakes, a concept quite alien to a machine. Robots with advanced AI features are being developed to deal with a multitude of functions which, in the past, were the exclusive domain of human beings, from doing household chores to disarming unexploded bombs. They are touted as being available as companions who can respond to requests by human beings and even carry on a conversation.

So can we say that such machines can think? That to some extent depends on what we mean by ‘think’. Alan Turing, considered to be the father of artificial intelligence, answered by stating that we need not decide whether a machine can think, but only whether it can behave as a person with human intelligence. On a number of tests one can devise, these machines seem to pass this test to some extent.

A recent book, Superintelligence: Paths, Dangers, Strategies, by Nick Bostrom, emphasises the problems associated with the creation of supercomputers, which will be far more intelligent than mere human beings. The author believes that this “is quite possibly the most important and most daunting challenge humanity has ever faced”.

This warning has been echoed recently by no less an authority than Stephen Hawking who has stated that with the rise of AI, it is a question of when and not if such creatures will take over the world and make slaves of us all.

Hawking argues that AI has the capacity to grow exponentially, whereas the human mind can develop only slowly over decades. Human intelligence, as measured by tests like IQ, has shown a steady but very slow improvement over the past 100 years, a process which certainly cannot cope with the rate of growth of AI.

Some philosophers have argued that a machine, however intelligent, can never take over some aspects of brain activity, particularly self-consciousness, and therefore can never become a sort of human being.

John Searle, professor of philosophy at the University of California, Berkeley, in a recent article in The New York Review of Books (October 9), does not believe that it is possible for a computer to deal with what he calls “observer-dependent information”; i.e. that sort of information which needs human consciousness to make sense of. He believes that machines, however intelligent, cannot have consciousness and therefore ‘intent’. They will never want to do anything. The Deep Blue supercomputer beat chess master Garry Kasparov not because it wanted to beat him but because it was better at choosing the right move from its vast repertoire of available moves.

As in many other situations with science research, we can expect to be confronted with a number of increasingly complex ethical problems. While there is no doubt about the immense benefit computers have given to humanity over the past half-century, it would be well to keep in mind the undoubted negative aspects that have emerged over this time and will continue to worry us in years to come.

On the bright side, one could be cheered by the thought that throughout history, and particularly in the history of science, it has been shown time and time again that grand prophesies about future developments made by the most outstanding gurus have almost invariably been proved to be wrong.