We are told that the person who invented the word ‘robot’ was Czech playwright Karel Capek, in his 1920 play RUR. It is based on the Czech word ‘robota’, meaning ‘labour’. This was the Capek family’s second stab at describing man-machines/ his elder brother Josef had tried ‘automat’ in a 1917 short story. But that came to mean specifically a vending machine. Anyway, it was just a variation of the existing Greek-derived word ‘automaton’. So the younger Capek hit the mark, as a consequence of which ‘robot’ is all he is remembered for.
RUR stands for ‘Rossum’s Universal Robots’, which sounds quite friendly. But this is a science-fiction play and, guess what, things go wrong. The plot has similarities to the Philip K Dick story which became the film Blade Runner. In Capek’s play, the Rossum factory produces robots as a source of cheap labour. These are given consciousness by scientists, and then turn against their makers by destroying the entire human race, apart from the scientist who created them. But they have a limited life – they wear out after 20 years. And eventually they get so good that you can’t tell the difference between them and humans.
At one point in the play, a human character remarks that since the robots in production remember everything but can think of nothing new, they’d make fine university professors. These are not mechanical devices, by the way, more bio-engineered, fleshy ones. It is when they acquire the human characteristics of empathy and love that they in effect become human: two of these humanoid super-robots are sent out into the world to repopulate it, a new Adam and Eve.
We’ve been worried about what robots might do to us – or how humans might start to genetically engineer themselves to effectively become robots, as in Aldous Huxley’s Brave New World – ever since. It’s happening: in the US they pretty much breed giant basketball players, and I’m convinced that a sub-race of weirdly proportioned supermodels is coming off a production line somewhere. And so we have the famous Turing Test, devised by computer pioneer Alan Turing in 1950. Can a robot persuade us it is human by replying convincingly to our questions, appearing to be able to think?
Various experiments suggest this can now be done. Equally, some humans can convince people that they are robots – or rather, ‘chatbots’ as automated call-centre voices are called. Moreover, android-like robots are being produced which are physically convincing because of the way they imitate the ‘micro-movements’ of human beings. They fidget, scratch their noses and so forth.
This is surely the penultimate design challenge/ to put all this together so as to create machines we can routinely mistake for people. Thus far, convincing androids exist only in computer animation. I don’t think a Rossum’s robot is close yet, for three reasons: appearing to think is not the same as thinking; machines have a way to go before they can self-heal like humans; and what powers them? We can hardly make an electric car last 160km, or a laptop more than six hours. Think of all the energy a robot would need to move and think, all day long.
And the ultimate design challenge? We’ll need a foolproof way of telling the bods from the ‘bots. Or a lot of university professors are going to get very nervous indeed.