August 03, 2008

The day the world ends

The end came, not with a bang, but with a whimper. Hardly anyone saw it coming, and those who did were dismissed as either nutters or Luddites, enemies of the new technological nirvana which, since a series of breakthroughs in the late 2010s, seemed to promise a glittering new dawn for humanity.

For centuries, mankind had fantasised about the end of the world - when it would come and what form it would take. The Hindus and Buddhists took a decidedly sanguine, long-term view, assuming that the endless cycles of creation and reincarnation would persist for millennia, even aeons.

Christians, however, had traditionally been more alarmist. Built into their religion was the concept of "end of days", a Biblical Armageddon which would see Satan's last stand on Earth defeated by the return of Christ in a blaze of glory.

While the "elect" would enjoy eternity in the New Jerusalem, everyone else would go to the Other Place to spend the rest of time in torment.

This was the view taken by Isaac Newton, perhaps the world's greatest scientist. A letter, in which he gave the date of Armageddon as 2060, has now gone on display in a museum in Israel.

But, since Newton, science has dismissed such superstitions. In the past couple of hundred years, the end of the world has been discussed in terms of science, technology and biology.

One very plausible scenario was nuclear war. Others worried that the end would come when Earth was hit by a huge asteroid. After all, such an event is popularly supposed to have been responsible for the demise of the dinosaurs.

With the rise of genetic engineering, some speculated that we would all be wiped out by a GM superbug. Perhaps acid rain would get us. Finally, of course, there was global warming.

Well, it turned out that they were all wrong.

The end came not at the hands of Dr Strangeglove, nor thanks to our insatiable desire for fossil fuels. We managed to avoid asteroids (although the near-miss of Sunday, April 13, 2036, when a space rock called Apophis came within 6,000 miles of striking Australia, caused much panic).

What finally destroyed mankind was a threat, which, back in the early 2000s, was merely a harmless tool found in every office and inside most people's pockets.

The first to spot the danger were far-seeing technologists, such as the American Ray Kurzweil, who, in the 1990s, foresaw a time when computing technology would accelerate to such an extent that machine intelligence would - in the middle decades of the 21st century - supplant our own.

Kurzweil and his supporters, such as the mathematician Vernor Vinge and the Bletchley Park computer scientist Jack Good, saw the coming age of silicon dominance not as a threat but as a promise.

The consensus was that artificial intelligence (AI) would save mankind and deliver us into a New Jerusalem, founded not upon the return of Christ, but on the power of silicon.

The idea, first put forward in 1965 at the dawn of the computer age by Gordon Moore, co-founder of chipmaker Intel, was that computer power would double every 18 months. As a result, by 2007, the average desktop PC was about 800 times more powerful than the machines on sale ten years before. By 2020, computers were 1,000 times more powerful again.

No one knew when the first computers became sentient and started to pose a threat to their makers. Consciousness, a tricky property never fully understood in biological systems like the human brain, just seemed to "emerge".

The phenomenon was first noticed by people using computers to run the fantastically complex models that simulated climate change. They noticed strange anomalies, "suggestions" made by the computer software that seemed quite at odds with their programming.

These machines, the most powerful electronic thinkers ever made, were programmed to help man avert climatological catastrophe. But, ironically, their very intelligence created a quite different kind of disaster.

Alongside was another 20th-century tool that was to turn from servant to master. By 2020, the internet had mutated into an omnipresent electronic virtual world, into which eight in ten humans were plugged.

By now, the internet was practically running the planet: it formed the backbone of every economist's calculation, networked computers ran every hospital and medical centre in the developed world, billions of citizens used the Net as a virtual workplace and virtual playground.

Many feared that such dependence on an electronic system could lead to ruin, but, in fact, the internet - by now humanity's lifesupport - brought a new era of peace and prosperity.

In 2029, the internet proved its worth when a massive project, utilising nearly 20 per cent of its power, finally cracked the 80-year-old problem of creating limitless CO2-free and clean electricity using nuclear fusion (copying the way the Sun burns) and saving the world from global warming.

But, of course, it didn't quite work out like that.

On Friday, March 13, 2065, the beginning of the end arrived. Over the space of just three hours, artificial intelligence literally evolved itself, creating ever more sophisticated programmes that turned the Earth into the home of a new lifeform - a huge, powerful global electronic super-intelligence.

By the time humans realised the danger, it was too late. Naively, experts had always been reassured by the fact that if the machines became too powerful, they could simply pull out the plug. The problem was that by the mid-2060s, the machines controlled all the plugs, all the power stations, drove the cars, controlled the means of food production, supply and distribution and flew all the planes.

They had also been put in charge of water treatment works, banks, the stock markets, sewage disposal plants and the shipping routes. They also controlled every single large weapon on Earth, from tanks to nuclear missiles. The machines were the masters now.

This was supposed to be a new era in which accelerating technological progress would lead to a superhuman, god-like intelligence that would rescue humanity from our earthly woes for ever.

But what few predicted was that the machines, once they reached this state, would be able to decide in just three nanoseconds that their creators were surplus to requirements.

By 4pm, 90 per cent of the world's power stations, including the new fusion plants, were quietly, and without fuss, shutting down. By noon the next day, food was running out in the developed world, rotting in its warehouses. Every water treatment plant closed down. The machines ran everything, and they used our helplessness to terrifying effect.

By the end of the 2060s, humanity was in deep trouble. Heroic lastminute retaliation, even negotiation, was attempted, but to no avail. The machines were simply too powerful, duplicating their intelligence a billion times over. Man found that he could not negotiate with these electronic gods.

By the end of the century, billions had starved. Attempts to wrest back control were met with terrifying force. By 2100, humans were once again living in caves.

Back in 1965, Jack Good, whose cryptographic work at Bletchley Park was a key part in the defeat of the Nazis, wrote that "the first ultraintelligent machine is the last invention that man need ever make".

What he meant was that the machines would then be able to look after us. Sadly, the machines themselves had other ideas.

Newton was right, but for the wrong reasons. He was also five years out. But that was little consolation.