Just over ten years ago I wrote a book called March of the Machines. In the first draft I claimed that it would be quite a few years before we would see robots walking around reliably on two legs; it was an extremely difficult task. The mathematics involved proved that to be the case. There was nothing like it at the time. But before the book was published I had to carry out a swift piece of editing. Honda's P2 biped had appeared, calmly strolling around, even getting up and down stairs with relative ease.

Ten years on and we have a plethora of two-legged robots. Asimo is perhaps the best known, possibly because of its dancing abilities and friendly disposition, serving drinks at functions and shaking hands with celebrities. At the latest count there are, in fact, more than 100 Asimos - probably even twice that number if you put all the spare limbs together. Then there are educational robots and toy robots, all similarly strutting about. Last year we even ran a competition at the Science Museum in London for schoolchildren, who were asked to program a MechRC two-legged robot to dance to funky music and follow this up by completing an assault course.

But this raises the question: given where we were in the late 1990s and where we are now, where will we be in ten or even twenty years time? What will robots be capable of in 2030? Will we be able to trust an Asimo servant not to spike our drink? Indeed, will Asimo still be our servant or will the roles have been reversed?

By then we may well have come to and even gone past what is commonly known as “the singularity”. This means that as the pace of technological change keeps speeding up, it will soon get to a point where normal humans will not be able to keep up and will be “cut out of the loop”. At this stage it will then be intelligent machines, robots if you like, or upgraded humans, augmented by a merger with technology, who become the dominant life forms on Earth.

But is this actually a possibility or is it merely something from the latest Hollywood science fiction? After all, robots can't even play football yet and how could a computer appreciate the smell of a rose or understand a joke? And even so, if a robot ever got out of hand, well, we could always switch it off, couldn't we?

Unfortunately, these comforting thoughts are pure Hollywood. Our lives today are highly dependent on technology in everything we do. To control technology completely, you must be able to switch it off. Can we switch off the internet today, practically, realistically? Of course not. We must not think in terms of individual machines, but a network that controls and links a plethora of technology. An intelligent network to boot.

And as for all these human values that we think are so great - when an intelligent missile system is firing at us, if we tell it that it can't play football as well as we can, or can't smell a rose like we can, we cannot rely on this to swing things in our favour. The missile system is unlikely to say: “Yes, you are correct. I must wait until my footballing skills are as good as yours before I fire at you.”

We must also remember that machine brains in an intelligent network are different from human brains. First, human brains are not networked in anything like the same way. We know that human mathematical skills and memory capabilities are vastly inferior in performance: these are reasons for using computers today. We must couple this with the fact that, as humans, we think in only three dimensions (compared with multidimensional machine thinking) and have only five senses: it is estimated that humans sense less than 5 per cent of what is going on around them. Worst of all, humans still communicate using mechanical pressure waves - speech. Just how embarrassingly poor can they be?

The simple fact is that robots are different from humans both physically and mentally. Robots do not have to be like humans and copy humans in every respect before they can outperform humans. They do not have to be conscious like humans, they do not have to understand the nuances of the English language, they do not have to appreciate Beethoven in order to be better than humans where it matters.

For several years after the Wright brothers had first taken to the air there were those who claimed that they were not flying: what birds do is fly and this was very different. When Deep Blue beat Garry Kasparov at chess, Kasparov claimed that he was cheated; it was unfair that the computer was networked. Quite simply machines do things in a different way. They are designed for a purpose, to do things differently.

When the Aztecs and the Native Americans were defeated by Europeans it could be said that the “better” culture lay with the home teams. What the invaders brought with them though, apart from disease, was a vastly superior technology that the home teams didn't understand and never got to grips with. This technology was applied very often ruthlessly with little thought for conscience, moral upstanding behaviour or culture.

So think of yourself as an Aztec of today, faced by a networked robot invader whom you do not understand, who doesn't stick to your rules of engagement, who is far more powerful and who out-thinks you at every turn. So do you still want to take that drink from Asimo now?

But there again, why not plug yourself into the network and upgrade. Do you fancy a brain implant before it's too late? Who does this Asimo think he is anyway?

Kevin Warwick is Professor of Cybernetics at the University of Reading

Times Recommends

• Resist urge to throw voters red meat
• Slay sacred cows and be candid
• We cannot let these 100 deaths deflect us