Flesh and Machines: How Robots Will Change Us

  • Rodney A. Brooks
Pantheon: 2002. 288 pp. $26
Credit: DAVID NEWTON

As robots are made to be more like humans, and human organs are replaced by mechanical artefacts, will humans cease to be more 'special' than machines? A hurried reading of the dust jackets of these two books may give the false impression that they are ill-informed texts written to create fears about advancing bionic aliens mischievously created in frankensteinian laboratories. Actually, nothing could be further from the minds of the authors. Both books cover serious scientific and philosophical issues in the design of humanoid robots and the replacement or supplementation of organs.

Rodney Brooks, an influential robotician from the Massachusetts Institute of Technology, has been building what he calls 'artificial creatures' for some years. In Flesh and Machines he writes autobiographically of his flight from 'representational' artificial intelligence, in which the programmer endows the robot with a full description of its environment so that it might act intelligently. Brooks now takes an alternative 'situated' approach in which the robot is equipped with ways of reacting appropriately to its environment without the burden of referring all its actions to pre-digested models supplied by the programmer. Indeed, one of his early models was an insect-like hexapod called Genghis that reacted to the heat emitted by mammals in its vicinity. Later systems include the celebrated Cog, a waist-up robot that moves its limbs and eyes appropriately in the presence of a human. Another machine, Kismet, responds to humans with appealing facial gestures.

An evident weakness in this approach is that Brooks may need to return to some inner representation to introduce the contemplation that humans use a lot of the time, before reacting to their environments. This does not need to be done through pre-programming but in a brain-like fashion through a rapid evolution of, and learning in, an artificial nervous system. Brooks argues that this is just an elaboration of what he has achieved already, but this remains to be seen.

The book moves from an autobiographical to a predictive style in which visions of robots range widely from helpful house servants to agents on planetary missions. But my greatest enjoyment came from the end of the book, which is more philosophical. Brooks' philosophy is blunt: abilities that some hold dear as being specially human (for example, emotions and consciousness) are available to machines simply because humans are machines, albeit immensely complex ones.

He prepares a robust response to potential attacks by detractors who believe that the 'specialness' of human consciousness either lies in an as yet uncomprehended presence of quantum physical action in the brain, or is not available to physics at all. He accuses some philosophers of simply being afraid to contemplate the possibility of a conscious machine. According to Brooks, the adversaries will have to prove their points or admit that humans can be understood from physical control principles. Despite this, Brooks has his own little je ne sais quoi which separates the biological from the artefactual. The 'juice', as he calls it, that is missing is a proper mathematical understanding of complex biological mechanisms. He does not see this as an insurmountable difficulty, but rather as an achievable and exciting target for science.

The Body Electric: An Anatomy of the New Bionic Senses

  • James Geary
Weidenfeld & Nicolson: 2002. 256 pp. £20

The second book, The Body Electric, is a clear and readable report of the way that human sensory systems are being replaced or supplemented by technological artefacts studied in a variety of research laboratories across the world. Journalist James Geary tells an exciting story not only by reporting his in-depth interviews with key researchers, but also by asking the beneficiaries of some prostheses what this has done for them. People who have seen, heard and learned to speak for the first time are the best ambassadors for this science — their accounts detract from any false impression that humanity is being insidiously invaded by the silicon chip.

Inevitably, reportage can fall prey to some odd views held by those who are being interviewed. A British engineer, for example, insists that a radio chip he had implanted under the skin of his forearm to open doors and cause his computer to greet him is a major step towards the automatic transmission of emotions and, eventually, thoughts, between humans. Here the scientific challenges of encoding thought and emotion distributed over billions of brain cells are suspended, and would a chip in the pocket not have opened the door just as well?

By contrast, the rest of the book presents real scientific work on vision, hearing, touch and even taste. The question is always the same: how can a sensory faculty be restored to an individual who may have lost it? Often this needs an enormous amount of ingenuity, mainly because the human brain is adept at integrating the input from many sensory modalities. For example, in order to 'see', the brain integrates signals from the retina with signals from the oculomotor system that moves the eyes. The key question is whether what's left of the brain's apparatus when parts go missing will be able to accept and adapt to the retinal implant in order to bring to the perceptual mind some sense of seeing. Geary answers this by sharing the delight of a 63-year-old Belgian woman who, after having been blind for over 20 years, began to see dots of colours when her optic nerve was stimulated by the output of a small video camera.

Having dealt with the senses, Geary approaches the problem of artificial minds. He talks to some who insist that they can create a mind just by throwing neural networks together and letting the mind emerge, as well as to others who have more principled approaches along the lines of the work reviewed by Brooks. Geary, rightly, just reports, and leaves it to the reader to spot the outliers.