The Taste of Others

Why the Japanese want their robots to act more like humans

HER name is MARIE, and her impressive set of skills comes in handy in a nursing home. MARIE can walk around under her own power. She can distinguish among similar-looking objects, such as different bottles of medicine, and has a delicate enough touch to work with frail patients. MARIE can interpret a range of facial expressions and gestures, and respond in ways that suggest compassion. Although her language skills are not ideal, she can recognise speech and respond clearly. Above all, she is inexpensive . Unfortunately for MARIE, however, she has one glaring trait that makes it hard for Japanese patients to accept her: she is a flesh-and-blood human being from the Philippines. If only she were a robot instead.

Robots, you see, are wonderful creatures, as many a Japanese will tell you. They are getting more adept all the time, and before too long will be able to do cheaply and easily many tasks that human workers do now. They will care for the sick, collect the rubbish, guard homes and offices, and give directions on the street.

This is great news in Japan, where the population has peaked, and may have begun shrinking in 2005. With too few young workers supporting an ageing population, somebody—or something—needs to fill the gap, especially since many of Japan's young people will be needed in science, business and other creative or knowledge-intensive jobs.

Many workers from low-wage countries are eager to work in Japan. The Philippines, for example, has over 350,000 trained nurses, and has been pleading with Japan—which accepts only a token few—to let more in. Foreign pundits keep telling Japan to do itself a favour and make better use of cheap imported labour. But the consensus among Japanese is that visions of a future in which immigrant workers live harmoniously and unobtrusively in Japan are pure fancy. Making humanoid robots is clearly the simple and practical way to go.

Japan certainly has the technology. It is already the world leader in making industrial robots, which look nothing like pets or people but increasingly do much of the work in its factories. Japan is also racing far ahead of other countries in developing robots with more human features, or that can interact more easily with people. A government report released this May estimated that the market for “service robots” will reach ¥1.1 trillion ($10 billion) within a decade.

The country showed off its newest robots at a world exposition this summer in Aichi prefecture. More than 22m visitors came, 95% of them Japanese. The robots stole the show, from the nanny robot that babysits to a Toyota that plays a trumpet. And Japan's robots do not confine their talents to controlled environments. As they gain skills and confidence, robots such as Sony's QRIO (pronounced “curio”) and Honda's ASIMO are venturing to unlikely places. They have attended factory openings, greeted foreign leaders, and rung the opening bell on the New York Stock Exchange. ASIMO can even take the stage to accept awards.

So Japan will need workers, and it is learning how to make robots that can do many of their jobs. But the country's keen interest in robots may also reflect something else: it seems that plenty of Japanese really like dealing with robots.

Few Japanese have the fear of robots that seems to haunt westerners in seminars and Hollywood films. In western popular culture, robots are often a threat, either because they are manipulated by sinister forces or because something goes horribly wrong with them. By contrast, most Japanese view robots as friendly and benign. Robots like people, and can do good.

The Japanese are well aware of this cultural divide, and commentators devote lots of attention to explaining it. The two most favoured theories, which are assumed to reinforce each other, involve religion and popular culture.

Most Japanese take an eclectic approach to religious beliefs, and the native religion, Shintoism, is infused with animism: it does not make clear distinctions between inanimate things and organic beings. A popular Japanese theory about robots, therefore, is that there is no need to explain why Japanese are fond of them: what needs explaining, rather, is why westerners allow their Christian hang-ups to get in the way of a good technology. When Honda started making real progress with its humanoid-robot project, it consulted the Vatican on whether westerners would object to a robot made in man's image.

Japanese popular culture has also consistently portrayed robots in a positive light, ever since Japan created its first famous cartoon robot, Tetsuwan Atomu, in 1951. Its name in Japanese refers to its atomic heart. Putting a nuclear core into a cartoon robot less than a decade after Hiroshima and Nagasaki might seem an odd way to endear people to the new character. But Tetsuwan Atom—being a robot, rather than a human—was able to use the technology for good.

Over the past half century, scores of other Japanese cartoons and films have featured benign robots that work with humans, in some cases even blending with them. One of the latest is a film called “Hinokio”, in which a reclusive boy sends a robot to school on his behalf and uses virtual-reality technology to interact with classmates. Among the broad Japanese public, it is a short leap to hope that real-world robots will soon be able to pursue good causes, whether helping to detect landmines in war-zones or finding and rescuing victims of disasters.

The prevailing view in Japan is that the country is lucky to be uninhibited by robophobia. With fewer of the complexes that trouble many westerners, so the theory goes, Japan is free to make use of a great new tool, just when its needs and abilities are happily about to converge. “Of all the nations involved in such research,” the Japan Times wrote in a 2004 editorial, “Japan is the most inclined to approach it in a spirit of fun.”

These sanguine explanations, however, may capture only part of the story. Although they are at ease with robots, many Japanese are not as comfortable around other people. That is especially true of foreigners. Immigrants cannot be programmed as robots can. You never know when they will do something spontaneous, ask an awkward question, or use the wrong honorific in conversation. But, even leaving foreigners out of it, being Japanese, and having always to watch what you say and do around others, is no picnic.

It is no surprise, therefore, that Japanese researchers are forging ahead with research on human interfaces. For many jobs, after all, lifelike features are superfluous. A robotic arm can gently help to lift and reposition hospital patients without being attached to a humanoid form. The same goes for robotic spoons that make it easier for the infirm to feed themselves, power suits that help lift heavy grocery bags, and a variety of machines that watch the house, vacuum the carpet and so on. Yet the demand for better robots in Japan goes far beyond such functionality. Many Japanese seem to like robot versions of living creatures precisely because they are different from the real thing.

An obvious example is AIBO, the robotic dog that Sony began selling in 1999. The bulk of its sales have been in Japan, and the company says there is a big difference between Japanese and American consumers. American AIBO buyers tend to be computer geeks who want to hack the robotic dog's programming and delve in its innards. Most Japanese consumers, by contrast, like AIBO because it is a clean, safe and predictable pet.

AIBO is just a fake dog. As the country gets better at building interactive robots, their advantages for Japanese users will multiply. Hiroshi Ishiguro, a robotocist at Osaka University, cites the example of asking directions. In Japan, says Mr Ishiguro, people are even more reluctant than in other places to approach a stranger. Building robotic traffic police and guides will make it easier for people to overcome their diffidence.

Karl MacDorman, another researcher at Osaka, sees similar social forces at work. Interacting with other people can be difficult for the Japanese, he says, “because they always have to think about what the other person is feeling, and how what they say will affect the other person.” But it is impossible to embarrass a robot, or be embarrassed, by saying the wrong thing.

To understand how Japanese might find robots less intimidating than people, Mr MacDorman has been investigating eye movements, using headsets that monitor where subjects are looking. One oft-cited myth about Japanese, that they rarely make eye contact, is not really true. When answering questions put by another Japanese, Mr MacDorman's subjects made eye contact around 30% of the time. But Japanese subjects behave intriguingly when they talk to Mr Ishiguro's android, ReplieeQ1. The android's face has been modeled on that of a famous newsreader, and sophisticated actuators allow it to mimic her facial movements. When answering the android's questions, Mr MacDorman's Japanese subjects were much more likely to look it in the eye than they were a real person. Mr MacDorman wants to do more tests, but he surmises that the discomfort many Japanese feel when dealing with other people has something to do with his results, and that they are much more at ease when talking to an android.

Eventually, interactive robots are going to become more common, not just in Japan but in other rich countries as well. As children and the elderly begin spending time with them, they are likely to develop emotional reactions to such lifelike machines. That is human nature. Upon meeting Sony's QRIO, your correspondent promptly referred to it as “him” three times, despite trying to remember that it is just a battery-operated device.

What seems to set Japan apart from other countries is that few Japanese are all that worried about the effects that hordes of robots might have on its citizens. Nobody seems prepared to ask awkward questions about how it might turn out. If this bold social experiment produces lots of isolated people, there will of course be an outlet for their loneliness: they can confide in their robot pets and partners. Only in Japan could this be thought less risky than having a compassionate Filipina drop by for a chat.

 

Dec 20th 2005 | TOKYO
From The Economist print edition

As Paul Otellini prepares to take charge at Intel next week, is the giant chipmaker heading in the right direction?

WHEN it was confirmed last November that Intel’s new boss would be Paul Otellini, the firm’s chief operating officer and heir apparent for several years, nobody was surprised. The firm’s succession-planning is famed for being as clean and efficient as the factories where it makes its chips, and Craig Barrett, the retiring chief executive, who steps down on May 18th, has long seen Mr Otellini as his right-hand man. But Mr Otellini is Intel’s right-hand man in another sense, too. For he is the architect of the firm’s new strategy—a change of direction that Mr Otellini calls a “right-hand turn”. The world’s largest chipmaker now faces three big challenges—but Mr Otellini believes his plan can address all of them at once. Is he right?

The first challenge is technical. For years, Intel has consistently improved the performance of its chips by making them run at higher and higher clock speeds (measured in MHz or GHz). But it has now hit a wall. As chips get faster, they consume more power and generate more heat. It also becomes harder to keep all the parts of a chip marching in step.

Intel, along with its rivals, is embracing a new approach to chip design, in which performance is improved not through higher clock speeds, but by adding further processing “cores” to its existing chips. A “dual-core” chip can, at least in theory, deliver twice the number-crunching performance of a single-core chip, but at the same clock speed. In practice, the software on the chip must be rewritten to exploit multiple cores, but such software is becoming increasingly common.

The second challenge is that the personal computer (PC) market has matured. Intel still makes most of its money selling the processor chips at the heart of PCs. Mr Otellini’s predecessors, who ran the firm when the PC was in the ascendant, could rely on an expanding market to provide double-digit growth for Intel. Mr Otellini cannot, and must find new sources of growth, in the PC market and beyond it.

The third challenge is the growing competitiveness of Intel’s main rival, Advanced Micro Devices (AMD). In recent months Intel has suffered a string of embarrassments, some self-inflicted (such as the cancellation or delay of several new products), but others at the hands of AMD. Notably, AMD devised a clever way to enable chips to handle data in both 32-bit and 64-bit chunks, which improves their performance. Having spent years developing an entirely new 64-bit chip, Itanium, Intel was loth to undermine its prospects by adding 64-bit support to its 32-bit Pentium and Xeon chips. But last year it did just that, to remain competitive with AMD. As a result, Itanium is probably doomed.

Mr Otellini’s response to all of these challenges is the “right-hand turn”. First, Intel must change how it designs chips. As well as switching to a dual-core (and then multi-core) approach, the firm is starting to integrate other functions, such as security and networking features, on to its chips.

That points the way to the second part of Mr Otellini’s plan: “platformisation”. Rather than just selling processing chips to PC-makers, Intel intends to offer them entire “platforms”—bundles consisting of a processor, its ancillary chips and networking components, and the software needed to tie them all together. By doing this, Intel hopes to sell more components, thereby taking a larger cut of the selling price of each PC. It also hopes to boost demand by devising specific platforms for several promising new markets, such as home entertainment, mobile devices and health care.

This strategy, it hopes, will also enable Intel to outflank its rivals, which specialise in particular kinds of chips, such as processors (as in the case of AMD) or networking (Broadcom). PC-makers, goes the theory, would rather buy a single integrated package from Intel than assemble components from several other suppliers.



A man with a plan
This all sounds good in theory. But will it work in practice? When asked, Intel executives invariably cite the success of Centrino, the firm’s laptop platform, which combines a processor with a Wi-Fi networking chip, software and other supporting components. Centrino, launched in 2003, is the model for how Intel intends to sell chips in future. Rather than ask for “Intel inside”, laptop buyers can now demand “Centrino inside”. This encourages PC-makers to buy the entire platform.

Intel is now working on two similar platforms for home and office desktop machines, nicknamed “Desktrino”, due to be launched later this year. In January Mr Otellini reorganised Intel into platform-specific divisions, including digital home (for consumer PCs), corporate (business PCs and servers), mobility (laptops and mobile devices) and health care, which Intel regards as a promising new market.

Will it work? Consider Centrino. When it was launched, several laptop-makers initially turned their noses up at Intel’s Wi-Fi chip, and decided to buy only Intel’s laptop processor chip, which they combined with Wi-Fi chips bought from other vendors. Only when Intel’s Wi-Fi chip came up to scratch did laptop-makers opt for the whole Centrino package. “It took a while for the Centrino model to work,” says Dean McCarron of Mercury Research, a market-research firm. The platform model will only succeed, he says, if all the components are competitive in their own right.

While Intel has a good chance of getting the platform model to work in desktop PCs, breaking into new markets may prove much harder. Intel has yet to make any headway in the mobile-phone market; indeed, merging its laptop and mobile businesses into a single “mobility” division means that continuing losses in communications chips are helpfully obscured by the bumper profits from Centrino. Its prowess in processors is unquestioned, but Intel is still to prove itself when it comes to radio chips. It must do so if it is to realise its high hopes for new markets such as smartphones and wireless broadband (Intel is the main cheerleader for WiMax, a new wireless broadband technology). “They don’t have a significant technological advantage over the incumbents in the cell-phone business,” says Kevin Krewell of Microprocessor Report, an industry journal. “Until they come up with something truly unique, they’re just going to be a me-too player.”

There are other worries too. Dell, the world’s top PC-maker and one of Intel’s closest allies, has hitherto chosen not to use AMD’s chips. But Kevin Rollins, Dell’s chief executive, admitted in February that his firm came close to changing its mind last year, following product delays and other problems at Intel. Dell is careful never to rule out abandoning its Intel-only policy. If Dell ever changes its mind, Intel’s market share in PC processors, which has exceeded 80% for years, could come under threat. Intel rewards Dell and other PC-makers with generous discounts, in the form of marketing subsidies, to keep them loyal. Dirk Meyer, the number two at AMD, says that Intel has a “stranglehold” over PC-makers. He points out that Japanese antitrust regulators recently ruled that Intel’s business practices in that country, which are similar to those used elsewhere, were unfair. Mr Krewell, however, says that Intel had worse practices in the past, and has cleaned up its act.

As Mr Otellini takes the helm, Intel is already starting to move in the direction he has charted. The origins of the new strategy can be traced back to a speech he made in late 2001. Since then he has been gently steering the firm into its right-hand turn. He is the first boss of Intel to have a business and marketing rather than a technical background (though, as a 31-year veteran at the firm, he is hardly a technological novice). As Intel faces the challenges not just of designing chips in new ways, but of selling them in new ways too, Mr Otellini would seem to be the right man for the job. He has mapped out where he wants to take the company: now he must get it there.

May 12th 2005
From The Economist print edition