by futurist Richard Worzel, C.F.A.
I’ve written about robots on several occasions before, but want to revisit the subject because the arrival of robots is imminent, and portends more than just science fiction characters come to life. First, let’s review where we are.
Robots have been around for decades, but mostly in very rudimentary forms, such as machines bolted to the floor of a car factory, or toys that roll over to you, carrying a tray. More recently, there have been what might be considered demonstration robots, such as Hitachi’s EMIEW, and Toyota’s Mobiro. As well, there are now commercial household robots, like Roomba and Scooba that, respectively, vacuum or scrub ceramic floors, as well as robots that pick up golf balls at driving ranges, and that mow lawns, both for individuals, and for industrial users. None of these have been terribly impressive -– but they weren’t really intended to be. Instead, they were stretching the possible, and testing out new techniques. The idea of a robot being able to look at the world, analyze its surroundings, and move on wheels, or walk on legs, evading obstacles, is actually a remarkably difficult problem. It’s one that computer scientists have been working on for decades – at least since I studied computer science in university in the early 1970s, but now have just about solved. Today’s robots walk on legs. The HRP-4C robot, built by Japan’s National Institute of Advanced Industrial Science and Technology, walks on two legs and has a human form. Big Dog, walks on four legs and looks like a headless mule. And the best way to view the development of robots – in all respects, not just in walking – is much as you would think of a human toddler: they are clumsy, and just learning how to walk, but they will develop quickly, now that they’ve taken their first steps.
Remember that the computers that direct robots will continue to follow Moore’s Law (“Computers will double in speed and halve in price every 18 months”), thereby becoming more powerful at exponential rates (or faster). Within a small number of years (somewhere between 2 and 5), we will see robots that can walk competently, and in a seemingly adult-human fashion and speed. And it’s not just the skill of walking that will develop: it’s robot’s overall abilities to perceive, analyze, consider, and respond. They will become progressively more and more life-like. Indeed, those who have seen the Big Dog video on YouTube have proclaimed it “creepy,” by which I infer they meant that it acts like a living thing.
There’s more, though: not only is the raw power of computing going to increase, some say by 1,000 times over the next 10 years, but there are many kinds of software that are being developed that allow computers to learn how to solve problems, and teach themselves new skills. The website shapingtomorrow.com cites an example of a Japanese robot that can create its own lab experiments, coming up with results unknown by its human creators. The combination of dramatically more powerful computers, and increasingly sophisticated software means that we are at the cusp of the emergence of robots.
And it’s not just robots: automation will get progressively more powerful and sophisticated, and, again, at exponential rates or better. This is both an opportunity and a threat. The opportunity is that increased automation means increased productivity – companies will be able to produce more, or do more, with the same number of people, or produce the same amount with fewer people. (This is, after all, what increased productivity means.) And increasing productivity, for an economy, generally means increasing wealth and affluence.
The threat is that automation is eating its way up the employment food chain, replacing workers with machines. This has been happening for centuries, starting with the Industrial Revolution, but the pace of change is now advancing at computer speeds, and the increasing sophistication of automated devices means that it’s not just grunt work jobs that are being eliminated, ringing up gas purchases, but jobs that require more, and higher, skills. Coming back to the Japanese lab robot mentioned earlier, suppose that within several years a robot can do the work of a skilled lab technician, one that trained for, say, three years, and holds a post-secondary diploma. What kind of work would such a highly trained individual look for then? She already has an undergraduate degree, plus a post-graduate diploma, and the automation replacing her will get smarter much faster than she can. What kind of future lies in store for her?
Which points to a broader question. Automation will continue to get better and smarter as far as we can foresee into the future, whereas humans are limited by the capacity of our brains and senses. When automation can do almost everything, what will there be left for humans to do? And what share of the higher standard of living will we be entitled to when we have done no work to contribute to it? Or will we become a society of owners and peasants, with the owners of the machines vastly rich, and the peasants, who do not own them, living off the crumbs of the aristocracy?
Ned Ludd, the leader of the luddite movement, had an answer: the luddites smashed the knitting machines that were producing better cloth, and faster, than a human, thereby displacing workers. Will there be a movement to smash robots and computers? That would be dangerous, because we’ve become so dependent on them, and will become more so in future.
And if we find a way to distribute the goodies that automation delivers, even to those who do not work, what then will we do with our lives? Is there life after work, or life without work?
But how soon is this likely to happen? Is 10 years wildly optimistic, or wildly pessimistic, or somewhere in between? No one knows, because there’s a lot of art involved as well as a lot of science. It’s not just enough that computer scientists can come up with the software and hardware to make these things happen. You also need the venture capital to back it, management to create a workable business plan to deliver it to market, and an education period for people to get used to the idea of robots and smart computers and get comfortable with them. If Big Dog, for example, looks “creepy” today, how long until “creepy” becomes “commonplace,” and acceptable?
Well, one of the best ways of gauging the progress of robots will be by watching the American military. The Predator unmanned (but human controlled) flying drones are now being widely used by the military, and to great (military) effect. Big Dog was developed specifically to help soldiers carry gear in terrain where wheels won’t work, pioneering a quadruped walking robot. DARPA, the Pentagon’s research arm, sponsored a prize that led to the development of the first self-driving car, a Chevy Tahoe, in a simulated urban environment, complete with traffic signals, other vehicles, and pedestrians. And you just know that DARPA has to be working on a robot soldier, to go to places too dangerous for humans. To the extent that they succeed (and it’s publicized), that will be the leading edge of the utilization of robots.
My own sense (and this is, at best, an educated guess) is that within 10 years time we will have robots in public and private places doing menial tasks, such as cleaning, washing, and general maintenance. (We already have robots like Roomba, a computer-run vacuum cleaner, so this isn’t much of a stretch.) Within 20 years, computers, robots, and automation will be seen as a serious threat to workers in many low-, medium-, and even some fairly high level jobs, and the use of computer-driven devices (whether robots or some other form of automation) will be a hotly debated political issue.
Yet, it may be that this happens just in time. In the developed world, the aging of the population is going to mean that, within 20 years’ time, there will be a lot of retired people who need help, and nowhere near as many who can give it. There are already villages in Japan that have no inhabitants under the age of 60, which is one of the reasons why Japan leads the world in robot research. It may be that the robots be the ones to do the work the boomers need done.
In other words, the robots may arrive just in time.
© Copyright, IF Research, June 2009.