Published: 24 Nov 2015
Reports of robots posing a threat to jobs are about as numerous as the ones that tell us to keep calm. Machines may do calculations swiftly, the soothsayers say, but they can't match humans' capacity for creativity.
Not yet, anyway. Columbia University's Hod Lipson gives it less than a century.
"What do people do, what happens to social structure, when machines can do most things better than most people?" the renowned robotics engineer asked at MIT's recent EmTech event in Cambridge, Mass. "This is not some hypothetical question about the future of AI [artificial intelligence] a million years from today. It's probably in the lifetime of our children -- it's something that we have to certainly start thinking about seriously."
CIOs are part of that we. Yes, today's chief information officers are probably safe from being replaced by automaton IT executives that are better, faster and cheaper -- even as companies invest in the next wave of digital technologies like cognitive computing. But advances in robotics, as Lipson noted, have implications far greater than the viability of any one corporate role.
Lipson works on "machine creativity," or endowing machines with the attribute human beings have long claimed as uniquely their own. He builds robots that do more than aggregate and analyze data; they generate "a lot of good stuff," including paintings and even other machines. But they're not simply equipped with high-tech printers or other tools; these are products of -- mirabile dictu -- imagination.
"If you ask almost any parent to talk about their child, if they talk about how bright their child is, usually they will talk about how creative and curious their child is," Lipson said. "We think that is something that's very, very human and very unique."
But creativity is notoriously difficult to teach even to humans. That's because we don't exactly know where our ideas come from. For machines, it's even harder, Lipson said, which is why AI has focused for so long on analysis. Computers can predict the stock market, the weather and what we're going to buy, but they can't come up with the idea for driverless cars.
Advances in robotics may change that. Lipson is taking the robot-building process out of human hands and turning it over to evolution, which he calls "the mother of creativity." Robot components like wires and motors are combined in a simulated evolutionary engine and piece themselves together. Robots are formed and tested, offspring are created through variation and crossbreeding -- and out comes a primitive-looking robot with tubular appendages and a finlike tail that helps it crawl through sand. Lipson hopes to move all the way up the chain, wherever that may lead.
That machines can come close to anything remotely resembling human evolution is bound to touch a nerve with a lot of folks. And rightly so. Lipson's success would mark the end of civilization as we know it -- never mind the supplanting of jobs that once provided comfortable middle-class lifestyles for millions.
In a 2014 Pew Research Center study, half of the 1,896 industry observers surveyed predicted that vast numbers of not only blue-collar manufacturing jobs will be taken in coming years by robotics and artificial intelligence but white-collar ones too. More recently, Forrester Research Inc. published a report that was generally sunny on the future job outlook. Still, it predicted that 22.7 million jobs -- 16% of the U.S. job market -- would be displaced by technologies, far outnumbering the 13.6 million they would create.
Meantime, the popularity of cognitive computing -- robot brains, if you will -- is surging. IBM recently founded a business unit to promote its supercomputer Watson as a cloud service, and research outfit IDC says more than 50% of developer teams will use some form of cognitive technology in their apps by 2018. In customer service, for example, cognitive systems can tap reams of customer data to answer questions in a conversational tone. In healthcare, they can spot genetic patterns in patients to help them make better decisions about their health.
And these are just apps. What happens when Lipson's evolutionary process produces a race of Renaissance robots?
"This is a forum for 'can we do it?'" said Jimmy Jia, an MIT grad and CEO of Distributed Energy Management who was in Cambridge for EmTech. The Seattle company he co-founded helps businesses manage their use of utilities like electricity and water. "Whether we should is another forum."
A place for us
For Lipson, there's no going back. Robots will get their share of creativity, and -- the hope is -- they'll use it to help humans come up with innovations they wouldn't have thought of otherwise.
"I think the real question in the long term is, 'How do we restructure our economy?'" he said. "We have to think about new ways of creating meaning and creating accomplishment through other things like art, sports, education -- things that are more transformative but are self-defined. That's a good place to look at."
He has a lot of company. In an open letter in the July/August issue of MIT Technology Review, MIT's Erik Brynjolfsson and Andrew McAfee, authors of The Second Machine Age, and a slew of others, including Salesforce's Marc Benioff, wrote that advances in robotics won't lead to a jobless future -- if governments forge the right policies. They need to make increased investments in education, infrastructure and R&D. And business leaders have no small task either, they wrote. They must reimagine the goals of corporations they're guiding into the future -- from mere profit centers to civic institutions that embrace "inclusive prosperity."
Other pundits call for guaranteed income for all as a way to combat a future where robots do all the work. CIOs, that's probably not something that's going to win you plaudits in the C-suite. But as you push toward a digital future, as you should, think of your children. And think of your children's children.
Advances in robotics means smart machines get smarter
Robots taking your job? Nonsense, said ComputerWeekly's Bryan Glick
Have a look at 10 of today's robots