Is
Einstein Robot too human? | Robot
wars of the future | Japan:
nation that loves robots | Calls
for 'Model-T robot' | Blog:
the robot receptionist | Comment:
tomorrow's dominant life-form
I was looking forward to my conversation with Ramona, a solid-looking brunette
who reminded me of Jemima Rooper, the leading actress in Lost in Austen. One
blogger said he'd been “blown away” by her. And so, at her behest, I
introduced myself.
Ramona: Hello, I'm Ramona. What's your name?
Me: I'm Anj. How's the weather?
Ramona: Hi there, the weather. What did you do last weekend?
Ramona is a chatterbot, a computer program that can supposedly hold a
humanlike conversation, an important aspiration for those working in
artificial intelligence. Her creator, the futurologist Ray Kurzweil, has
written a film around her in which she tries to obtain full legal rights as
a person but proves to be more clever than the human beings judging her.
Judging by our exchange, I'd say Kurzweil is a bit of an optimist. Ramona's
inability to respond authentically to my fairly easy questions shows that we
don't need to worry about being usurped by artificial beings just yet. And
when you try to capture the essence of humanness in a robot, things look
even more pear-shaped. Although Honda spent 20 years teaching Asimo, its
advanced humanoid robot, how to walk, it still looks slightly constipated.
Sebastian Thrun, director of Stanford University's Artifical Intelligence
Laboratory, notes that “[the field of] robotics is incredibly young. We are
in the 16th century of robotics right now. And computer science is young -
maybe in the 17th century.”
But in laboratories all over the world, engineers, technologists and futurists
persist in trying to replicate the wondrous panoply of powers possessed by
people, either their physical abilities (doing the housework,
admininistering medicines, driving) or their intellectual capacities (such
as language, ability to play chess, or perception of facial expressions).
Most successful robots do only one thing; we are still a long, long way -
possibly centuries - from bundling both the physical and mental agilities of
humans into the scary amoral android so beloved by science-fiction writers.
No matter; some robots do their one thing terrifically well. An example is the
Einstein robot developed by David Hanson, at Hanson Robotics, which is
designed to detect and respond to an observer's facial expressions. Hanson,
a former Disney engineer, used a patented, skin-like polymer called Frubber
to construct his homage to the great man. The spookily realistic robot, with
trademark unruly hair and moustache, can smile, furrow its brow and
generally look responsive. In the absence of a clear practical application
for a stretchy, swivelling Einstein head, its most useful role might be to
determine whether we human beings want robots to look like us. “Some
scientists believe strongly that very human-like robots are so inherently
creepy that people can never get over it and interact with them normally,”
Hanson told a website recently. Reactions to footage of the Einstein robot
on YouTube have been mixed; many people confess to being unnerved by it.
Thrun, who is trying to develop an intelligent car that drives itself, makes
the same point: “I think we have to see whether there is a very strong
desire to replicate human behaviour and human looks...in the next ten or 15
years we will find out whether, for example, an elderly person would be more
accepting of a humanoid robot as a companion, or a robotic wheelchair or
robotic walker that doesn't look like a human but provides specific
functions.”
Most robots are far more mundane than the Einstein one and look nothing like
us. A washing machine is a robot of sorts - it performs, mechanically, a
task that human beings used to do. Even heavily publicised domestic robots,
such as the Roomba (a vacuum), aren't much to look at: it resembles a thick
Frisbee that lurks on the floor, waiting to be activated so it can swish
around sucking up dirt and navigating corners and obstacles (although it
can't dislodge itself if it gets stuck).
Robots are used routinely in industry, such as car assembly, and to carry out
tasks in extreme environments (think of robots that map the ocean floor,
travel to other planets or crawl into volcanoes). They tend to have narrowly
defined tasks and to be operated remotely by human beings. Some are equipped
to make limited decisions - such as which direction to move in - but they do
not possess what we consider intelligence.
There are attempts, though, to minimise the input of human operators,
sometimes for very good reasons. Last month, the Massachusetts Institute of
Technology unveiled an “intelligent” robotic forklift truck for use in war
zones; apparently, forklift drivers in Iraq have to run for cover around
four times a day. The truck is given a guided tour of its environment and
then, using voicerecognition software to “understand” what it's seeing, can
memorise the layout. After that, it interprets commands given remotely, much
as a soldier in the field will interpret the duties assigned to him by a
commanding officer.
Still, the truck is some way from being an autonomous agent capable of making
its own decisions based on what it senses around it. That is because it is
notoriously difficult to instil any kind of humanlike intelligence into
robots. For example, Asimo has only just been given the ability to avoid
bumping into people when it walks. Its head - basically a helmet - is
equipped with cameras and sensors that calculate the speed and distance of
approaching pedestrians and allows it to take evasive action.
It is this idea of autonomy that jolts us from our comfort zone. Investing a
machine with the power to act without human authorisation, without human
direction, seems alien to us. Yet some robots have been built as killing
machines. South Korea has developed $200,000 machinegun-toting robots that
it plans to station along its border with North Korea. Reports claim the
machines are programmed to distinguish between friend (eg, a human soldier)
and foe by asking for a password, and the failure to produce one could
result in the robot opening fire.
Such developments have prompted discussion of whether international bodies
should come together to devise a code of ethics that should pervade
robotics. The best known ethical code is that crafted by Isaac Asimov, who
came up with the Three Laws of Robotics. These are: 1. A robot may not
injure a human or, through inaction, allow a human to come to harm; 2. A
robot must obey orders given by a human unless these conflict with the first
law; 3. A robot must protect its own existence as long as this does not
conflict with the first or second law.
However, this assumes a level of intelligence in robots that is at least
decades away, if not longer. Noel Sharkey, Professor of Artificial
Intelligence and Robotics at Sheffield University, has identified a more
immediate worry. In a recent article for Science, he noted that 14 companies
in South-East Asia are developing robots for childcare and sounded caution
about leaving babies with robots for long periods. Studies have shown that
children are remarkably responsive to robots, and one experiment even showed
that toddlers could be persuaded - groomed?- to treat one like a friend
rather than as a toy.
Sharkey wonders whether parents might be tempted to leave their children in
the care of a robot such as PaPeRo, made by the Japanese firm NEC (which
does not explicitly advocate such a use, although it markets the cute,
dancing, singing PaPeRo as a “childcare robot”) and says that long periods
without human interaction can be psychologically damaging. He believes that
there is a need to discuss the ethics of using robots with vulnerable
groups, such as children and the elderly: "We were caught off guard by
the sudden increase in internet use and it would not be a good idea to let
that happen with robots,” he says.
It wouldn't, not if Ramona's off-the-wall conversational style is anything to
go by. Here is what happened when I tried to bid her farewell.
Me: Goodbye Ramona. I have to eat now.
Ramona: Oh, don't fret, that's a condition that is easily curable.What's going
to be your first course of action in ridding yourself of this illness?
Me: Hunger is not an illness
Ramona: All right. So where do you meet people? I spend most of my time on the
net, so I'm a bit out of touch
Me: You should get out more.
Ramona: That's an idea, but I'm rather shy.
Me: Nonsense. You're very friendly. Gotta go. Bye.
Ramona: OK, thanks. Maybe you have other suggestions?
Me: Nope.
Ramona: I understand it very well.
Me: Goodbye.
Ramona: Until next time the weather.
Robots through the ages
1495: Leonardo's robot
Leonardo da Vinci sketched a humanoid figure designed to be able to wave its
arms, move its head and open its mouth. The “robot”, in his diaries, was
encased in a suit of armour.
1937: Elektro
Built by a US electricity company, Elektro was 7ft tall, capable of speech
(through a record player) and could walk. He was a heavy smoker.
1969: Shakey
Developed at Stanford Research Institute, California, Shakey was the first
robot capable of detecting its environment.
1997: Sojourner rover
This six-wheeled robot analysed rock samples from the surface of Mars and sent
550 photographs back to Earth during the Pathfinder mission.
2002: The Roomba
This robot is designed to do one duty: vacuuming. More than two million units
have been sold since its introduction in 2002. Although it has special
attachments to tackle pet hair and will chirrup triumphantly when it has
finished cleaning a room, the Roomba is not designed for tackling shagpile.
2007: The BEAR
A prototype under development by the US robotics company Vecna, with US Army
funding, the Bear is designed to find, lift and transport injured soldiers.
Vecna's website says: “Whether on a battlefield, in a nuclear reactor core,
near a toxic chemical spill or inside a structurally compromised building
after an earthquake, the BEAR can rescue those in need without risking
additional human life.”
Six robot witticisms
“Do you want me to sit in the corner and rust, or just fall apart where I'm
standing?”
Marvin, the Paranoid Android (The Hitchhiker's Guide to the Galaxy)
“It's the old story: droid meets droid, droid becomes chameleon, droid loses
chameleon, chameleon turns into blob, droid gets blob back again, blob meets
blob, blob goes off with blob and droid loses blob, chameleon and droid. How
many times have we heard that story?”
Kryten (Red Dwarf)
“Will Robinson! I will tell you a joke! Why did the robot cross the road?
Because he was carbon bonded to the chicken!”
The Robot (Lost in Space, 1998)
“Oh my goodness! Shut me down. Machines building machines. How perverse.”
C-3PO (Star Wars) on witnessing a droid assembly line
“My worst nightmares are being realised. I can see me now, tin cans, TV
dinners, garbage can lids.”
K.I.T.T. (Knight Rider) on finding himself in a car compactor
“Blackmail is such an ugly word. I prefer 'extortion'. The 'X' makes it sound
cool.”
Bender (Futurama)
