¿ìè¶ÌÊÓƵ

Wasubot the robot is a musical prodigy. At the tender age of five it
played the organ in concert with one of Japan’s leading orchestras. A few
years later, equipped with more sensitive mechanical fingers and a new name,
WABOT-2, it was wowing audiences with piano recitals. Now, at the dizzy
peak of its musical fame, it can contemplate a future line of descendants
leading much more serious lives. Thirty years from now robots modelled on
WASUBOT and WABOT-2 could be whizzing around Japanese hospitals helping
to fetch, carry and nurse the sick, or helping elderly people fend for themselves
at home.

In Japan, the era of the silicon serf may be just around the corner
if you believe Ichiro Kato, one of Japan’s leading robotics engineers. For
the past decade, Kato and his team at Waseda University in Tokyo have been
winning over sceptics by gently nurturing the musical skills of robots like
Wasubot, refining the sensors they need to respond intelligently to sound
and touch. Last year the research entered an ambitious new phase, ‘project
humanoid’. For the next nine years, the Waseda engineers will strive to
turn their musicians into agile, intelligent robots with a streak of human
kindness. The research is deadly serious. It promises the ultimate in high-technology
solutions – robotic slaves – to the time bomb ticking at the heart of Japan:
an unprecedented increase in the number of old people .

All the world’s industrialised nations are seeing increases in the proportion
of old people. But Japan’s population is changing fastest of all. A postwar
leap in life expectancy and a plummeting birth rate mean that the proportion
of Japanese aged 65 or over will almost double over the next three decades,
reaching 25 per cent of the population by the year 2025. That percentage
will not be much different in Germany, Sweden and Britain. But unlike Japan,
the shift will have happened gradually, accumulating demographic change
over a century or more.

‘Unlike Western European countries, we don’t have the luxury of time,’
says Makoto Atoh, director of the Institute of Population Problems in Tokyo.
Early next century Japan’s labour force will start to shrink just as the
number of elderly people needing nursing rockets. Already Japan has some
400 000 bedridden senior citizens and 800 000 victims of strokes or Alzheimer’s
disease. By 2025 these figures will have tripled. ‘The critical period is
between now and the year 2000,’ says Atoh. ‘That’s the time to prepare.’

If history is any guide, the solutions Japan comes up with could end
up as exports to ageing industrial rivals. With daunting optimism, Kato
sees the personal robot as the model-T Ford of the 21st century. Japan,
he predicts, will move seamlessly from a ‘motor car economy’ to a ‘computer
economy’ to a ‘humanoid robot economy’. Even if his bold vision crumbles,
the quest for more automation won’t. Japan will persist with less ambitious
technologies aimed at reducing labour costs in its health-care system.
Automated machines for dispensing drugs and satellite link-ups are already
creeping into its hospitals. More are in the pipeline.

Yet few policy makers pretend that labour-saving technology holds all
the answers to demographic upheaval. A host of social and economic measures
will be just as crucial, says Atoh. These range from raising retirement
ages, building basic nursing homes and reforming Japan’s complex health
insurance system. On all these fronts, Japan lags woefully behind its industrial
rivals.

A few decades ago people died young in Japan, and those who didn’t
were cared for by their children. As a result, the health-care system evolved
to cater for people with acute illnesses rather than the chronic conditions
associated with old age. Doctors and hospitals, for instance, recoup their
costs from the insurance system by prescribing drugs or performing clinical
tests, not by providing basic nursing care. The result is that Japan has
the highest per capita drugs bill in the world, and an embarrassing shortage
of beds in nursing homes. Often, elderly people needing basic nursing care
are admitted to hospitals where they are subjected to a barrage of unnecessary
tests and medication.

In a desperate bid to fix all this before the demographic avalanche,
Japan’s Ministry of Health and Welfare is hoping to build thousands of new
nursing homes and care houses – smart short-stay homes where elderly people
can be nursed back to health and, in theory, returned to their homes. But
high land prices – up to 2 million yen (about £12 5000) for an area
the size of a bed – mean that progress so far has been slow: only two new
care houses have been built in Tokyo. And in the meantime, two in three
existing nursing homes teeter on the brink of bankruptcy.

Add to this list of woes an explosion in labour costs next century,
and the enthusiasm for robotics is easier to understand. Optimists like
Kato predict that primitive humanoid robots will be assisting nurses and
technicians in hospitals within fifteen years. A couple of decades later,
he says, more sophisticated humanoids will enter old peoples’ own homes,
giving them the extra support they need to continue working, or helping
their younger relatives who care for them to hold down their jobs. There
is nothing heartless about this vision, he insists: ‘Elderly people would
find themselves more at ease with a personal robot than being a burden to
their families.’

Role reversal

Speaking from a tiny office in an engineering department that, disappointingly,
looks more like a workshop for Thomas the Tank Engine than a set from BladeRunner,
Kato spells out his philosophy. So far, the history of automation has been
about humans adapting to machines. In future, the pattern must be reversed
because ‘elderly people are not so good at manipulating machines’. To Kato,
accomplishing this reversal means only one thing: friendly anthropomorphism.
‘If it doesn’t walk and act like a human, it isn’t a robot – it’s merely
²¹³Ü³Ù´Ç³¾²¹³Ù¾±´Ç²Ô.’

Cynics take heed. With its camcorder head and ungainly frame, WASUBOT
might look like a sorry parody of a fictional robot now but the excellence
of its engineering makes Kato’s future look more plausible. Each hand has
21 degrees of freedom and each leg 4 degrees of freedom. And its 50 pneumatic
joints are controlled by 80 microprocessors, connected in series and in
parallel. ‘It flinches when you touch it with something hot,’ says Kato.
Moreover, the microprocessors are arranged in three hierarchical layers
so as to mimic the motor system of humans.

For all this, teaching WASUBOT the piano was far from easy. Organ keys
act as simple switches but the piano demands key strokes of varied pressure.
‘We added two degrees of freedom to the thumbs, giving each a total of four,’
explains Kato. Pressure sensors on its fingers tell the microprocessors
when to adjust the velocities of the key strokes. WASUBOT can strike a
piano key 15 times a second, twice as fast as most professional pianists.

Dazzling dexterity, though, is not the only reason the Waseda researchers
see WASUBOT as a good springboard for developing humanoid nurses. The robot’s
real strength, says Kato, is its ability to adapt itself to humans, albeit
in a primitive way. When it hears a human singing voice it will attempt
to accompany it on the keyboard, selecting the correct score from a database
of songs. If the robot’s initial key is outside the singer’s range, WASUBOT
will switch to a more conducive one. And if the song is not in its repertoire,
its camera eye simply needs to scan the score for a few seconds to store
it in its memory for future recitals.

Kato is the first to admit there is long way to go. WASUBOT’s fingers
might respond to pressure but they cannot yet extract information about
the texture of a surface. And no one is claiming to know how to build the
brain of a humanoid robot. Simulating all the functions of the 20 to 30
billion nerve cells that make up a human brain is out of the question. But
breakthroughs in nanotechnology might bring simpler artificial nervous systems
within reach, argues Kato. ‘After all, who would have predicted forty years
ago that a computer the size of a building could be shrunk to a single chip?’

If he is vindicated, the attempts to breathe life into WASUBOT won’t
stop with a thinking computer brain. ‘Domestic robots would also have to
be compassionate,’ says Kato, ‘and compassion does not exist without emotions.
In the case of developing robots for elderly people, we must install some
kind of capacity for emotions.’ Like Dorothy’s Tin Man, WASUBOT needs a
heart.

Fanciful as artificial emotions may sound, Kato is not the only researcher
entertaining such thoughts. Rodney Brooks, a robotics expert at the Massachusetts
Institute of Technology, and his colleagues have begun a long-term project
to develop a human-like robot. At the moment they are concentrating on
building the head and torso. But eventually the MIT team will try to simulate
emotions. Nor do philosophers and artificial intelligence experts necessarily
see any fundamental obstacle. ‘It makes all the sense in the world to think
machines could experience emotions,’ says Daniel Dennett, a philosopher
at Tufts University and a leading writer on consciousness. ‘But in the
end it might be better to keep robots grimly impersonal. There’s nothing
more off-putting than phoney anthropomorphism.’

If and when humanoid nurses, with or without emotions, reach the wards,
the chances are that hospitals will already have been revolutionised by
less daring technologies. Japan is not the only industrialised nation trying
to develop ‘smart’ hospitals as havens of information technology, ‘doctorless’
wards and electronic monitoring. Nor is it the only one dabbling in ‘telemedicine’,
an emerging practice in which general practitioners, hospital consultants
and patients interact through high-definition TV monitors. But faced with
the prospect of such rapid demographic change, Japanese reserchers are
keener than most to embrace these labour-saving remedies.

With its grim facade and peeling paintwork, Tokyo University Hospital
looks anything but smart. Yet plying its shabby corridors are small fetch-and-carry
machines, programmed to take linen and surgical goods to operating theatres.
And at the hospital pharmacy, automatic drug dispensers handle many prescriptions.
‘It’s not possible to automate the dispensation of all drugs because they
vary so much in type and shape,’ says Shigekoto Kaihara, chief of the Division
of International Health at the hospital. ‘But for tablet-type drugs the
approach is useful.’

There are moves afoot to go much further. In recent years, Japan’s bureaucrats
have invested millions of yen in projects involving medical checkups by
cable TV, plastic cards carrying digitised medical records and electronic
surveillance of hospital patients. The philosophy is simple. As biosensors,
optical fibres, and TV cameras and monitors improve and become cheaper,
there will be less and less need for doctors and patients to meet. Elderly
or disabled people living in remote rural areas could have their weekly
checkups without leaving their homes. And hard-pressed hospital doctors
could keep tabs on pulses, haemoglobin counts and glucose levels without
hiking over to distant wards.

But turning this dream into a reality could prove hard work. A seven-year
pilot project on remote monitoring of patients, set up by the government
in a hospital in Chino in the countryside of central Japan recently reported
mixed results. Using TV monitors, cameras and existing cable TV connections,
the researchers linked the hospital to a local nursing home and to the homes
of elderly, disabled people. The biggest problem, said one doctor, is colour
control. ‘Diagnosing even a simple condition like anaemia requires an accurate
knowledge of the patient’s colour. The image isn’t always accurate.’

Uphill struggle

Other difficulties include money and safety. ‘Who would meet the running
costs or take responsibility in the event of an accident?’ asks Kimiya Kato
of the Medical Information Development Center, a quango set up in the 1980s
to deal with automation in health care.

Doctors and policy makers are more confident about using TV systems
to transmit medical images to hospitals for specialist advice. Bringing
health care to Japan’s remote southern islands is difficult and costly because
they are spread over such a large area. But the doctors working there benefit
from being able to transmit X-rays and high-definition TV images of patients
to a large central hospital. ‘Telepathology’, as the practice is becoming
known, is also in full swing in hospitals in Tokyo and Sendai, where consultants
routinely make clinical decisions based on TV images.

With all the technological will in the world, however, Japan still faces
an uphill struggle in maintaining its health system and economy next century.
Robots are fine in principle but, in the short term at least, are unlikely
to undercut the workforce of nurses and care workers from Japan’s poorer
neighbours in Asia.

The very idea of immigrant labour strikes an unpalatable chord in Japan.
Yet the country is already home to an estimated 400 000 illegal aliens.
As one of a small group campaigning for ‘guest’ visas that would guarantee
such workers legal protection against unscrupulous employers, Atoh is in
no doubt about their importance. ‘The reality is that even with better technology
and robots we will still need foreign labour.’

* * *

Not so booming babies

As the populations of Africa and India explode next century, Japan will
be grappling with the opposite problem – a dearth of babies. Its present
birth rate, 1.53, is well below those of other industrialised nations,
and many Japanese demographers predict it will fall even further before
the end of the century. If the birth rate stays level, Japan’s population
will be halved by the end of the next century.

Why is parenthood so unfashionable? The short answer, says Makoto Atoh,
director of the Institute of Population Problems, is a surge in the proportion
of single women in their mid to late 20s. The fertility rate for couples
has changed little over the past decade, he says, but the birth rate as
a whole has fallen because women are choosing on average to marry later
in life. At between 26 and 27 years, the average marrying age for Japanese
women is now nearly two years higher than that of women in England.

Atoh attributes this gap to profound changes in the ‘status, role and
expectations’ of Japanese women. ‘Japan is experiencing the kind of social
upheaval that Britain and the US experienced in the 1960s.’ Education is
one of the main engines of change. ‘Women are entering universities in increasing
numbers,’ says Atoh, and the proportion choosing to study social sciences,
law and engineering – as opposed to arts courses – is rising. ‘In the past,
entering college was seen simply as a passport to marriage and a good husband.
Now women want a passport to a career.’ Atoh’s research shows that the more
education a woman receives, the more likely she is to remain single.

The 1980s economic boom made getting that career all the easier. But
the rest of Japanese society failed to keep pace. The structure of families
and the domestic expectations of men are much the same today as they were
a decade ago.