Welcome to the new Independent website. We hope you enjoy it and we value your feedback. Please contact us here.


Electric dreams: Is it the end for robot development?

We were promised a life of leisure thanks to hard-working robots and fiendishly clever cyborgs. But the android fantasy has largely been terminated, argues Michael Fitzpatrick

For years technocrats have been touting robots as the next big revolution, so big that their importance will rival that of car production and that they will create a utopia, "an Athens without the slaves", as British agriculture minister Peter Walker put it in 1983. But despite such claims, the development of mass-produced useful, commercial robots hasn't much moved beyond the glorified nut tighteners that work at car factories. Some think it never will.

The two countries that produce most of the world's robots, and dream of substituting immigration with metal people, Japan and South Korea, have been particularly gung-ho about this supposedly new lucrative market. A few years ago, Korean experts predicted a robot for every home by 2010. Obviously none of this has transpired, yet the robot industry and the technocracy continue to hype the robotic dawn as something akin to the second coming. Even Bill Gates believes the industry "is developing in much the same way that the computer business did 30 years ago". Nothing could be further from the truth.

After years of grabbing headlines with PR stunts involving walking, talking humanoids such as Honda's Asimo, Japan has at last conceded that perhaps what were needed were not more human-like robots, at least not in the next five years, but robots along the lines of the best-selling low-tech, robotic vacuum cleaner Roomba. This realisation seems to have dashed cold water on decades of technical arrogance.

Last year a Japanese government body, the New Energy and Industrial Technology Development Organisation, announced more funding for the development of less-glamorous robots and has earmarked 7.6bn yen to get these more prosaic drones and lifters into our homes. Until now, the emphasis in Japan was on creating humanoids and billions has been thrown at developing multitasking human-like robots that offer very few practical applications, either now or in the near future.

"They should be able to do more," says Joseph Engelberger (pictured above), the founding force behind industrial robots and considered the father of the modern robotics industry. "We need multitasking robots that can think for themselves and do something useful. Working robots have to be something more than this," he says, referring to the impracticality of most robots, at least as far as the media's opinion goes.

Now in his 90s, Engelberger sees the attempts to build human-like bots as something of a red herring and far from desirable. "The Japanese and Koreans grow up with a vision of robots are 'friends' so like to anthropomorphise them, which leads them to skirt the practicalities," he says. "The Japanese like to put a face on things, to make them look like humans or animals. It's more done for entertainment value than real practicality."

The Japanese Robot Association predicts that next-generation robots will generate up to $64.8bn (£42.8bn) in production and sales and $21.6bn in applications and support. However the same organisation recently pointed out that shipments of industrial robots actually fell 33 per cent in the last quarter of 2008, and 59 percent in the first quarter of 2009. Statistics from the British Automation and Robot Association show that, since 2000, the annual number of industrial robots installed in British businesses has fallen from a high of nearly 2,000 to fewer than 800.

Even taking the recession into account, demand for robots has been dramatically down worldwide, putting a severe dent into the dream of sparking another industrial revolution. Could it be that a flesh and blood employee is still far in advance of most bots in terms of capability?

"The more you to do with robots the more you realise just how good humans are," says Geoff Pegman, the managing director of one of Britain's few robot manufacturers, R U Robots. "It's difficult to instruct robots, especially in the office environment. For example, you can get them to deliver coffee, but that's it," he says. Fortunately, the slow progress in bringing intelligent mechanical help into the home, the office and even the factory (welding, yes. Building a carburettor, no) means the doomsayers who have been foretelling the collapse of civilisation and the dawn of human obsolescence in the face of the machine have also got that completely wrong.

"Mass unemployment? They said the same about the computer – how we would all have nothing but leisure time as computers would be doing our jobs. And this obviously is not the case," points out Pegman. And while the prototypes are becoming more physically capable, robots are still failing when it comes to the type of human common sense that allows us to walk around a corner without having a confidence crisis.

"Without 'consciousness', robots are just spot welders, bolt-tightening arms, burger flippers, spray painters, and factory drones that can swivel dexterously back and forth in pre-programmed for assembly line work. They're expensive toasters – one-trick-ponies. It will be many decades before anthropomorphic robots are going to be sentient," says Taro Hitachi, a Tokyo-based Japan blogger and an expert in patents. "Decades of creating fantasy robots, like Asimo, Aibo, Roborior et al, has bankrupted the research and development departments of Japan while not resulting in any viable products."

Nor does he accept that Japan's high robot count is a true reflection of the reality: "Japan counts almost any kind of semi-autonomous factory machinery as a 'robot', whereas in the rest of the world, milling machines and potato pealing machines are just factory automation."

Meanwhile, those that we all recognise as proper, autonomous, humanoid robots are finding their services hard sell to even in robot-friendly Japan. The country's biggest robot maker Tmusk created the lifelike Wakamuru robot five years ago, which it has pitched variously as a hospital porter, a receptionist and, most recently, a decidedly wooden actor, but has struggled to find interested clients. Costing £65,000 a piece, a rental program was scrapped recently because of lack of interest. And now thanks to the robot's less that scintillating talents, Tmusk's multibillion-yen helper robot project is mothballed.

Likewise, Honda has been working on robots since 1986, but finding it hard to make any money from its efforts. Culminating in the Asimo humanoid (its name a play on the Japanese word for "legs"), it first became available for rental in 2000. There has been no serious commercialisation of Asimo since then and the droid still needs a small army of helpers whenever it performs, although it has grown smaller and lighter over the years.

So why do big corporations still insist on throwing money at what Engleberg derides as "toys"? According to robotics expert Lem Fugitt, the Japanese are culturally predisposed to liking, even preferring, humanoid robots. Secondly many still believe, as he does, that robots could save the country from a fast approaching manpower shortage. He says: "Based on their Shinto and Buddhist heritage, the Japanese tend to believe that most objects incorporate some spirit and feel a connection with robots that move, that is, are animated. They believe that robots, especially humanoid robots, have many things in common with human beings."

Brought up on amicable robots such as Astro Boy, Japan is more ready to accept robots than the west. One survey has shown that 83.6 per cent of Japanese thought they would be able to live with robots.

Fugitt also points out that the biggest problems facing the country is a rapidly growing senior population with the longest longevity of any nation, and a shrinking number of active people in the workforce. Robots, he believes, are expected to take up the slack.

"The solution of both of those problems requires the development of advanced robotic systems, preferably in humanoid form. For the ageing population, it means developing healthcare and nursing robots," he says. For the shrinking work force it means supplementing human workers with robots capable of taking responsibility for more complex and demanding tasks than conventional factory automation robots have done in the past.

Despite numerous false predictions like this from the past, robot fans still cling to the idea that machines can be a practical aid around the home and even in hospitals.

"We think our robot will help make up for future labour shortages in an aging society with fewer children," said Osamu Tsuchikura of Fujitsu's robotics division shortly before his department was closed for good. Instead, dreams of domestic robots have been eclipsed outside of Japan by a sudden surge in the supply of a surplus of cheap, educated labour that instead continues to do the hard work that was once considered likely to become robots' domain. What could be automated is done manually now – call centres are just one example. But importing cheap labour as other nations, including Britain, have done to address the problem of labour shortages is not an option for Japan or South Korea. In fact Japan seems dead against liberalising its immigration policy. But as the robots needed for the work never materialised, Japan now faces sharp industrial decline.

The influx of cheap labour elsewhere, meanwhile, means putting the brakes on technological development of robots in the workplace. Why waste billions on research when there are cheap, skilled immigrants to do the job? The only area where robotic research is moving along rapidly is in the armaments industry – witness the "drones" that the US is using in Afghanistan and Pakistan. We may never see a friendly R2-D2 cooking our suppers in any of our lifetimes – but a Terminator, human shaped or otherwise, is marching towards the frontline.