FEATURE ROBOTIX 96
'Vvireé
wh
. §.0.0$':
.50
and thT one
.I.
. A...
..~..
‘ .t ..
a :q '. . ,4. .
0'0 n. a.“ y g ’ ‘0’. :U _» o'/ .-’ '
Computer boffins from all over Britain will be bringing their robots to Glasgow in an attempt to encourage public debate about our technological future. So will it be silicon chips with everything? wonders Eddie Gibb.
t the start of the movie RoboCop, the sales reps from a new-tech company proudly unveil their law enforcement android, armed to the teeth with machine guns which will enable it to clean up the streets in double-quick This being a futuristic nightmare. inevitably the control software goes haywire and the recalcitrant robot begins shooting the clients.
Granted it’s not much of a sales pitch for the future, but Dr Gillian Hayes of Edinburgh University’s artificial intelligence department found herself sympathising with the more mundane problem the android had in negotiating the stairs. For Hayes. or anyone working in robotics, this is literally the kind of stumbling block they come across every day.
After the success of Glasgow’s Robot Olympics in 1990, held to show that science is as much a part of culture as art, a two-day event called Robotix 96 is being staged, allowing the public to compare the fact and fiction surrounding our mechanised offspring. The robots are coming, for one weekend at least, and the word is they won’t look much like 3 Terminator or C3PO or K9 for that matter. It appears that anthropomorphic androids will be confined to Hollywood movies for some time to come.
In Edinburgh’s A.I. department — one of only two in the world — finding a definition of intelligence is of far more interest than what
time.
1‘;The List 8-21 Mar 1996
shape the hardware is. A group of post-graduate students has recently taken delivery ofGillespie — so-called because it resembles a pint of stout on castors — which is an off-the-peg robot to be used as a research tool exploring how machines can be made to learn. By posing the question ‘can machines be made to think‘. the research group hopes to provide some answers about the nature of human intelligence. A.l.. it seems. is closer to psychology or even philosophy than
‘The idea was that we would have lots of spare time as robots filled the factories. It would be a kind of steel- collared workforce to liberate us, but in fact it’s quite the opposite.’
the number-crunching and code-writing of computer science. Teaching a computer to beat a Grand Master at chess is seen as less worthwhile than understanding the process by which computers learn.
Both Hayes and her PhD student Simon Perkins believe that, eventually. machines will think like humans but that day is a very, very long way off. Before then, they have to teach Gillsepie how to climb stairs and avoid bumping into things. That in itself could represent a Iit’etime’s work. Ask not what a robot can do for you; how can it help you understand yourself is the more interesting question, according to Hayes. ‘What we’re doing is biologically inspired — it’s not engineering,’ she says.
The merging of biology and robots is the central theme of British cyberpunk wn'ter Simon lngs’s work. In his latest book Hotwire he coins the term ‘datafat’ to describe an imaginary organic compound which allows computer chips to be hooked up to the brain. Not so far—fetched, he says. considering heart pacemakers are already an accepted fact of life. Few people consider that as tampering with God’s grand design, but the idea of computer implants in the skull seems to be crossing some kind of imaginary moral line. That line was crossed long ago. argues lngs, who points to brain-altering drugs from aspirin to Prozac or the use of life- support systems in hospitals as evidence that humans are already happy to tamper with what is assumed to be the mind.