Virtually alive

First let us postulate that computer scientists succeed in developing intelligent machines that can do all things better than human beings can do them. In that case all the work will be done by vast, highly organised systems of machines and no human effort will be necessary. Either of two cases might occur. The machines might be permitted to make all of their own decisions without human oversight, or else human control over the machines might be retained (by an elite).

"If human control over ... machines [is] retained ... And if the elite [who control them are] softhearted liberals ... they may decide to play the role of good shepherds to the rest of the human race. [But] life will be so purposeless [for the majority] that they will have to be biologically or psychologically engineered to make them ... harmless. These engineered human beings may be happy in such a society, but they most certainly will not be free."

A man who could write these words would, to many people, be a fantasist who mistook science fiction for reality. And at the very least, their author - Theodore Kaczynski, otherwise known as the Unabomber, a former university professor who tried to murder several American scientists - was unbalanced in mind.

But Ray Kurzweil takes Kaczynski's speculations - and some even more startling than those - very seriously. And Kurzweil is no lightweight or loony. He's credited with the first print-to-speech reading machine for the blind and other devices for the disabled, and has contributed to other significant advances in Artificial Intelligence. MIT, the world's leading technical university, has honoured him as both Inventor and Engineer of the Year.

Dr. Kurzweil argues that technological development will proceed at such a breakneck pace in the next few decades that before the end of the next century, humanity will, for all intents and purposes, have merged with its (Information genetics and nano-) technology and have moved beyond biology as we know it.

He bases his case on what he terms a universal law of accelerating returns, found in certain kinds of highly complex systems. Information technology, it's argued, provides a prime example of this. Here it's knows as Moore's law, whereby integrated circuits double in complexity every year to eighteen months (a phenomenon first observed in 1964). By 1988 computer memory cost one hundred millionth of what it did in 1950. Since then, processing sped has continued to increase exponentially. In 1999, US$1,000 buys computing power roughly equivalent to the brain of an insect. Extending the trend, by 2009 a $1,000 personal computer (in 1999 dollars) will be able to perform about a trillion calculations per second. By 2019, the same money will buy a computing device with the massively parallel computational ability of a human brain (10 to the 15 calculations per second - or ten million billion).

Some philosophers, neuro-scientists and others have taken strong exception to what Kurzweil entertains as real possibilities. An example is the "inevitability" of machines becoming conscious and ultimately much cleverer than human beings. These critics highlight profound differences between the brain and mechanical processors. Certainly, such differences have been readily apparent with the technology developed up to now. Some thing computer scientists thought would be simple - such a distinguishing one face from another - turned out to be extremely complex; while solving some mathematical puzzles which are fiendishly difficult for people is easy for computers. Individual connections in the brain are very slow compared with those in silicon (in part because so much of the brain's energy is devoted to keeping the living substrate alive); but the brain makes up for it by its massive parallel capacity which makes possible ten million billion calculations a second (to which one is tempted to add, not in my head there aren't). By contrast, until fairly recently, massive parallel processing was making little progress in mechanical computers.

But others say Kurzweil is right on target. Among the arguments that lend his position credibility is the fact that, relentlessly, problems identified as insurmountable barriers to computers acquiring human abilities - or more - are being broken down. Beat the world champion at chess? No problem. Neural nets and evolutionary algorithms are another instance. Quantum computing? Yes, computers will be able to do that too.

In such a context, pronouncements about absolute difference begin to look as quixotic as Goethe's claim that a small bone in the human ear supposedly possessed by Man but not by other animals proved our species was altogether distinct. An another part of Kurzweil's future world, nano-engineering (the design and manufacturing of machines and other objects based on the manipulation of atoms and molecules), which was not far from science fiction ten years ago, is now taking shape as an infant industry in the United States, already attracting millions of dollars investment.

Environmentalists, then, would do well to take this book seriously, even if not literally. Information technology has long been recognised as a two-edged sword by the environmental movement. On the one hand, it has been an enormous benefit in campaigns - for example in the use of the Internet against the Multilateral Agreement on investments. "If a negotiator says something to someone over a glass of wine" boasted Maude Barlow, chair of the council of Canadians in 1997, "we'll have it on the Internet within an hour, all over the world." In April 1998, the OECD announced a six-month delay in negotiations, acknowledging that the NGOs had aroused enough opposition in many countries to derail the process. It was the Net, above all, which made this possible.

On the other hand, of course, information technology makes possible an ever greater acceleration of the globalised economy that many hold responsible for the worst aspects of environmental destruction and social injustice. As if to reinforce the point, neural nets and evolutionary algorithms came into commercial use in international investment funds in 1998.

Writing in an often jocular tone, Kurzweil does what he can to tackle the heavy moral and philosophical issues surrounding what he seems to see as a steady march forward: When do we consider machine intelligence to be conscious?: ("When the machines tell us so, and we believe them"); Whither nature? ("The laws of physics are not repealed by intelligence, but they effectively evaporate in its presence.") At times, these salvos go far afield, as in his belief that "we" will achieve a state of virtual immortality. But, for the most part, he is content to play the role of agent provocateur: "for now, it's enough just to ask the right questions." He is also well aware of some of the likely downsides - particularly the emergence of hugely sophisticated bioweapons - and the (to my mind mostly nasty) possibilities of 'virtual sex'.

But there are many issues he does not address, and campaigners will have to do a lot of the thinking themselves. If Kurzweil is right, the combined effect of the changes he predicts will, at the very least, completely redraw notions of scarcity and the value of human labour in industrialised economies. Who will profit, politically and financially? What will this mean for the protection and conservation of the non-human living world? In a world where the distinction between real and the virtual erodes - at least in over-charged brains in technology-obsessed cultures - will nature lose out even more?

On the downside, both the 'virtual war' veneer on the real-life-flesh-and-blood tragedy in the Balkans, and those insane teenage killers in Colorado point to what a "melding" with technology might do to the human spirit. It's clear that over-exposure to television, a very primitive technology by comparison with that Kurzweil predicts, can have devastating effects. A respected large-scale study on the impact of TV on violence published in the Journal of the American medical Association found that everywhere TV is introduced there is an immediate explosion of violence among young children in the playground, and within 15 years there is a doubling of the murder rate. (Why 15 years? That is how long it takes for the brutalised three-to-five-year-old to reach "prime crime age.")

And the upside? For one thing, superior computing power will bring vastly enhanced capacity to bear on modelling and understanding of potential challenges such as climate change. But one word you will not find in Kurzweil's book is hubris, and Spiritual Machines should carry a health warning to the effect that it should only be ingested through a filter of knowledge of the ancient Greek tragedies. In this case, as in so many others, an observation made well over a hundred years ago by Friedrich Nietzsche hold good: "Technology is the premise whose thousand-year conclusion mankind has not yet dared to draw."