The robot is a doctor or lawyer rather than hairdressing

Ava seems a bit helpless to look around. The robot, about a meter high, has a screen as face showing the person that her (or him?) Drives remotely.

Those can Ava virtual drive through a building and talk via built-in microphone and speaker with passersby. She finds her way largely independent. Convenient to attend remote meetings for example.

Now they pass through a room in a suburb of Boston full prototypes, naming the manufacturer iRobot cool stuff room. But Ava suddenly freezes at the door and bounces a few times with its wheels against the rim. That's pretty hard: the glass rebate vibrates it. "Open the door, please," appears on Ava's screen, she turns vaguely around. Robot confused. "A glass door is one of the hardest things to recognize the visual systems for robots," explains product manager Marcio Macedo of iRobot, which with around half a billion euros in sales and 530 employees is the world leader in commercial robots. "Just as birds for example, may have trouble with windows."

It's clever in that a device can navigate independently through buildings. But the clumsiness with glass doors of Ava, one of the most advanced robots of its kind, contrasts with the debate raging worldwide about robotics. What kinds of jobs are like robots steal? That question is at the top of the agenda for many entrepreneurs, researchers and policy makers, including the Netherlands. Next month presents the Scientific Council for Government Policy, a report about it. The concerns are about more than just the traditional, upright robots like Ava, but also on artificial intelligence and software robots are getting smarter.

What is often missing in these discussions: how hard it is going to do with the technology itself? If a door continue too difficult, what robots can then really any better than humans?

"Uhm .. Welding" said Daniela Rus. She is professor of robotics and director of the robotics institute of American technological university MIT in Cambridge, near Boston. That is one of the most advanced laboratories in the world. "Welding is difficult and dangerous for humans, robots do it better and faster." Even with heavy lifting and other repetitive physical tasks robots beat people with ease. "But lay a table or dishwasher make room, they still far. And you should not ask them to tie "Eye-hand coordination shoelaces, fine motor skills in changing situations, a space mapping and run through it as if humans can:. They're bad in there.

Spectacular plansIMPORTANT ROBOT HUB

For robots you have to be in Boston. The university town has emerged in recent years as one of the main spots for robot development.

Not only is the Massachusetts Institute of Technology there with a robot laboratory. Also iRobot, the world leader in household robots comes from.

Large companies such as Amazon and Google have established their robot divisions. And lately creates all kinds of robot startups, such as Re / think Robotics and Jibo.

Boston is not the only robot hub. Also in Japan, South Korea and Silicon Valley industry is growing rapidly.

That will still take a while, thinks Chris Jones, director of technology development at iRobot."Hardware is hard, it's just very hard to get all locomotion, sensors and visual systems well into line with one another." Many competitors get iRobot lately the news with spectacular plans and slick crowdfunding campaigns for advanced robots. The company Boston Dynamics flaunts example showing films of superstrong robots effortlessly over hills running back. But there are little to no robots are actually on the market that can do more than one specific, repetitive task. "For millions of dollars to make a prototype is slightly different than an affordable mass product," says Jones. He does not make any specific prediction."But we are still a long way. That will take many years, perhaps decades. "

Illustrates how difficult it is with the hardware is a YouTube video which went viral this summer. It shows a compilation of all collapsing robots on the DARPA Robotics Challenge, one of the most important events in the robot world. All kinds of robots tripping, falling over edges and lose their balance at demonstrations, even knulliger than the other.

Artificial intelligence

But before they can be completely derided as clumsy, there's more to robots than the hardware. Software, for example. Also thanks to the cloud, remote servers that robots are connected, a growing power that robots have at their disposal hard. The 'brain' of many robots consists of systems based on artificial intelligence. This is for instance the software can recognize data patterns in large amounts, and remembers those patterns so that he gets smarter by itself. The more computing power has such a system at his disposal, the faster the development. "It's really exciting that it's happening now," said Professor Rus.

Perhaps the best known example of a system based on artificial intelligence is Watson, from automation IBM. Of course, like IBM itself feeds wild futuristic visions of what later can do.But what Watson can now really?

"Actually, it's better than expected," said Mark Kris, oncologist at Memorial Sloan Kettering Hospital in New York. He gives that prestigious research hospital since 2012 led a project in which he makes better tailor cancer treatments for patients with Watson. Watson plowing through all the latest medical research and discovered therein patterns and connections.Which then he links to the medical histories of patients and other factors to determine what treatment the doctor should choose specific types of cancer. There is enormous variety in fact.

As an intern

"Now you have to Watson still learning all just like an intern. But in some areas it has already experienced scholar colleague who very much knows "Watson may not be perfect:. That human doctors because either. But at the same time are to be taken large bumps, for example, the delivery of structured data. For an excellent working Watson have excellent patient information required, which is now often lacking. "IBM's goal of a device that knows everything and maybe even take independent decisions about treatments, is still far away.Decades I think. But the initial results are promising. "

Watson sets, as competition from other companies, even though his first steps in the legal profession. A robot will not soon hold a flaming plea; creativity and social skills needed for this, understand artificial intelligence yet. But in the entire Dutch Code discover just that legislation which is relevant to a particular case, it all can be better. Much better than the average trainee lawyer likely.

Robot Software therefore progressing faster than robot hardware. If you do relates to jobs, which would logically mean that thinking is adopted rather than fine motor work that matters. Robots are previously lawyer or doctor or hairdresser.

That just lawyers and doctors should pay attention to robots, is also the point of the recently published book The Future of the Professions of Oxford economists and Daniel Richard Susskind. Not only because the technique itself and because it sure lends his subjects of choice where pattern recognition is important. But also because the professions in which it pays to automate tasks, if only because of the hourly rate of the people who practice it.

Really his doctor

Oncologist Kris does not think that doctors will ever be completely automated. Robots and computers will particularly ensure that physicians much more forward to seeing patients, he thinks. "Watson can take over a part of the diagnostic and administrative burden so the doctor just get time to really be a doctor."

There is also scientific evidence that says Kris. Last summer the Central Planning presented a study which called job polarization turns. By robotics and automation are changing the quality jobs it is true, but they persist. Like the crafts in the lower segment. But jobs are high speed automated.

Look at what is happening with banks and insurers: that automate recent years tens of thousands of jobs, especially at secondary level. Think of customer service representatives and order processors. If your work consists only of discovering patterns in data, sitting behind a computer and clicking a mouse - than it is to watch over the coming years.

Greater inequality

That is also the message of Andrew McAfee, co-author of the bestseller The Second Machine Age. His ideas have great influence worldwide, including the Netherlands. About half week he goes to a congress of this newspaper in discussion including minister Lodewijk Asscher (PvdA) of Social Affairs. McAfee is generally positive about robotics, but warns of greater inequality.

The core of his argument is that developments are going faster than ever, even faster than people can comprehend. All sorts of technological developments, in recent years, namely mutually reinforce one another: the cloud, artificial intelligence, sensors. With better sensors get better data, allowing artificial intelligence in turn is faster clever example. Moore's Law states that computers every two years, twice as fast. There does an unprecedented number of companies to robots, which allows for quick cuts.

An example that McAfee and his supporters often cite is the autonomous car. Thus it has gone much faster than experts thought a short time ago. Until a few years ago many thought that for decades would be far away. Also self-propelled cars were making until recently easily clumsy movies, but last week was starting manufacturer Tesla working self drive functions on the highway. As entire industries bets on robots such as the automobile industry did in recent years, can go fast, even if the technique is not so far.

Time for children

There is Professor Rus of MIT Robotlab totally agree. "But you must remain cautious with such predictions. I am optimistic. To begin with, there is no such thing as a fixed amount of work: a robot learns something, it means that the person who first made again to do something else useful. And those are often nicer things. For example, I really like a robot to which mine was folding, so I have more time to spend with my children, or nice work. "

She points out that technological development so far has been very good for the welfare of people: who has just always created more jobs and more prosperity. "And moreover, robotics is not a light switch that goes on or off. It is a gradual process in which there is still time to prepare for it. "The fact is that the disappearance of jobs due to technological development is already foretold centuries, but never really became reality.

Chris Jones of iRobot also warns against too outlandish predictions about robots. "The expectations of the public are sometimes too high. Robots are science, not science fiction. "There has to do a lot of hard work before robots really massive job can take over. And even if that happens, according to him no cause for fear. "Robots are going to help people, not replace it."

Wouter van Noort/ nrc.nl/handelsblad/Illustratio Rik van Schagen / Studio NRC