Extract of a speech, ‘The World in 2050 and Beyond’, by Lord Rees, Astronomer Royal and member of the Science Museum Group Foundation, at the inauguration of the Hans Rausing Lecture Theatre, in which, among other topics, he looks at the rise of robots and AI.
The smartphone, the web and their ancillaries would have seemed magic even 20 years ago. So, looking several decades ahead, we must keep our minds open, or at least ajar, to innovations that might now seem science fiction.
There’s been exciting advances in what’s called generalised machine learning. The London-based company DeepMind last year achieved a remarkable feat, its computer beat the world champion in the game of Go, and Carnegie-Mellon University in Pittsburgh has developed a machine that can bluff and calculate as well as the best human players of poker.
Of course, it’s 20 years since IBM’s ‘Deep Blue’ beat Garry Kasparov, the world chess champion. But Deep Blue was programmed in detail by expert players. In contrast, the machines that play Go and Poker gained expertise by absorbing huge numbers of games and playing against themselves. Their designers don’t themselves know how the machines make seemingly insightful decisions.
The speed of computers allows them to succeed by ‘brute force’ methods. They learn to identify dogs, cats and human faces by ‘crunching’ through millions of images, not the way babies learn. They learn to translate by reading millions of pages of, for example, multilingual European Union documents (they never get bored!).
But advances are patchy. Robots are still clumsier than a child in moving pieces on a real chessboard. They can’t tie your shoelaces. But sensor technology, speech recognition, information searches and so forth are advancing apace.
They won’t just take over manual work, indeed plumbing and gardening will be among the hardest jobs to automate, but routine legal work (conveyancing and suchlike), medical diagnostics and even surgery.
Can robots cope with emergencies? For instance, if an obstruction suddenly appears on a crowded highway, can Google’s driverless car discriminate whether it’s a paper bag, a dog or a child? The likely answer is that its judgement will never be perfect, but will be better than the average driver. Machine errors will occur, but not as often as human error, but when accidents occur, they will create a legal minefield. Who should be held responsible – the ‘driver’, the owner, or the designer?
The big social and economic question is this: Will this ‘second machine age’ be like earlier disruptive technologies, the car for instance, and create as many jobs as it destroys? Or is it really different this time?
The money ‘earned’ by robots could generate huge wealth for an elite. But to preserve a healthy society will require massive redistribution to ensure that everyone had at least a ‘living wage’, and to create and upgrade public-service jobs where the human element is crucial and is now undervalued and demand is huge, especially carers for young and old, but also custodians, gardeners in public parks and so on.
But let’s look further ahead.
If robots could observe and interpret their environment as adeptly as we do they would truly be perceived as intelligent beings, to which (or to whom) we can relate. Such machines pervade popular culture; in movies like Her, Transcendence and Ex Machina.
Do we have obligations towards them? We worry if our fellow-humans, and even animals, can’t fulfil their natural potential. Should we feel guilty if our robots are under-employed or bored?
What if a machine developed a mind of its own? Would it stay docile, or ‘go rogue’? If it could infiltrate the internet, and the internet of things, it could manipulate the rest of the world. It may have goals utterly orthogonal to human wishes, or even treat humans as an encumbrance.
Some AI pundits take this seriously, and think the field already needs guidelines, just as biotech does. But others regard these concerns as premature and worry less about artificial intelligence than about real stupidity.
Be that as it may, it’s likely that society will be transformed by autonomous robots, even though the jury’s out on whether they’ll be ‘idiot savants’ or display superhuman capabilities.
There’s disagreement about the route towards human-level intelligence. Some think we should emulate nature, and reverse-engineer the human brain. Others say that’s as misguided as designing flying machine by copying how birds flap their wings. And philosophers debate whether ‘consciousness’ is special to the wet, organic brains of humans, apes and dogs, so that robots, even if their intellects seem superhuman, will still lack self-awareness or inner life.
And now a digression into my special interest, and Hans Rausing’s: the cosmos and space. This is where robots will surely be transformative.
During this century, the whole solar system will be explored by flotillas of miniaturized probes, far more advanced than the robot that ESA’s Rosetta landed on a comet, or NASA’s New Horizons probe that transmitted amazing pictures from Pluto, 10,000 times further away than the moon.
These two instruments took ten years on their journeys, and the amazing Cassini probe of Saturn is even more of an antique – it was launched 20 years ago. Think how much better we could do today.
And better, too than the ‘Curiosity’ rover on Mars.
Later this century giant robotic fabricators may assemble vast lightweight structures in space, gossamer-thin radio reflectors or solar energy collectors, for instance, using raw materials mined from the Moon or asteroids. But what about human spaceflight? Robotic and AI advances are eroding the practical case.
Nonetheless, I hope people will follow the robots, though it will be as risk-seeking adventurers rather than for practical goals. The most promising developments are spearheaded by private companies. Elon Musk’s Space X, has launched unmanned payloads and docked with the Space Station, and has successfully recovered and reused the launch-rocket’s first stage presaging real cost-saving. He hopes soon to offer orbital flights to paying customers.
Wealthy adventurers are already signing up for a week-long trip round the far side of the Moon, voyaging further from Earth than anyone has been before. I’m told they’ve sold a ticket for the second flight but not for the first flight.
We should surely acclaim these private enterprise efforts in space, they can tolerate higher risks than a western government could impose on publicly-funded civilian astronauts, and thereby cut costs compared to NASA or ESA. But should they be promoted as adventures or extreme sports, the phrase ‘space tourism’ should be avoided. It lulls people into unrealistic confidence.
By 2100 courageous pioneers in the mould of say Felix Baumgartner, who broke the sound barrier in free fall from a high-altitude balloon, or Sir Ranulph Fiennes, may have established ‘bases’ independent from the Earth, on Mars, or maybe on asteroids. Musk himself (aged 45) says he wants to die on Mars, but not on impact.
But don’t ever expect mass emigration from Earth. Nowhere in our Solar System offers an environment even as clement as the Antarctic or the top of Everest. It’s a dangerous delusion to think that space offers an escape from Earth’s problems. There’s no ‘Planet B’.
Indeed, space is an inherently hostile environment for humans. For that reason, even though we may wish to regulate genetic and cyborg technology on Earth, we should surely wish the space pioneers good luck in using all such techniques to adapt to alien conditions. They’ll be free from terrestrial regulation and have maximal incentive to do so. Indeed, these spacefarers may spearhead the post-human era, evolving within a few centuries into a new species.
The stupendous timespans of the evolutionary past are now part of common culture, outside ‘fundamentalist’ circles, at any rate, but most people still tend to regard humans as the culmination of the evolutionary tree. No astronomer can believe that. Our Sun formed 4.5 billion years ago, but it’s got 6 billion more before the fuel runs out, and the expanding universe will continue, perhaps forever. To quote Woody Allen, “eternity is very long, especially towards the end.”
So, we may not even be at the half-way stage of evolution.
It may take just decades to develop human-level AI, or it may take centuries. Be that as it may, it’s but an instant compared to the cosmic future stretching ahead.
Moreover, the Earth’s environment may suit us ‘organics’, but interplanetary and interstellar space may be the preferred arena where robotic fabricators will have the grandest scope for construction, and where non-biological ‘brains’ may develop powers than humans can’t even imagine.