I found this while browsing through the Guardians site last night and thought it was rather interesting to think that they hadnt already started a
project/program on robots that can think and act for themselves.
But then again we never really know what any of our governments are up to, so they might have already done this and not said, which wouldnt be
anything unusual would it?
Will Fact Match Fiction As Scientists Start Work On Thinking Robot?
Alok Jha, Science Reporter
Monday August 25th 2003
Scientists have been given the biggest ever grant to build a "conscious robot". The work will not only bring the scores of intelligent, self-aware
machines that populate science fiction a step closer, it could also provide valuable clues on how human consciousness develops.
"Consciousness is perhaps the last remaining mystery in understanding what it is to be human, said Owen Holland, who will lead the work at Essex
University. By attempting to build physical systems which can produce a form of artifical consciousness, we hope to learn more about the nature of
To those outside the field, modern robotics may seem something of a damp squib. Robots might well have revolutionised manufacturing and taken over
many of the mundane, repetitive tasks once carried out by people, but the mechanised arms and small talking dogs built today are a far cry from the
sophistication of Douglas Adams' Marvin or Schwarzenegger's terminator.
The reason modern robots do not fire the imagination is that, at the moment, they are only capable of following instructions: they have no capability
to "think" for themselves. This is where the 500,000 (pounds) award to researchers from the universities of Essex and Bristol (UK) comes in.
Mr Holland's idea is that a consciuos robot would have to build up intermal models - one for the "self" of the robot and another for the world
Humans do this kind o evaluating almost without thinking. In order to be able to plan realistically, for example, we need to know not only our
physical limitations are (what we can do) and what is the best choice if we consider our options (what we should do) but also what we are likely to
choose (what we would do).
"My analogy is the recovering alcoholic who would like some cigarettes" said Mr Holland, "He can go to the bar next door or the tobacconist half a
mile away. If he fools himself that that he can just go into the bar and just but the cigarettes, (he knows that) he will buy a drink. In order for
him to plan successfully, he needs to know what he would do".
These models would be the basis from which the robot would build experiences. By experimenting with how its own body reacts with the world around it,
the robot will learn what is beneficial to it and what is not.
But does working out how to adapt to its environment actually make the robot conscious? Here the robotics researchers enter thorny philosophical
territory - what exactly is consciousness?
In an attempt to cut through the multitude of conflicting ideas and give the discussions more scientific rigour Igor Aleksander, an emeritus professor
of neural networks at Imperial College, recently described what he called"five major qualities" that are needed to form consciousness in living
beings and, subsequently, in machines. True to scientific form, th idea is that people can study and experiment with the axioms and propose any
The key to consciousness could lie in the robots models of itself. "I came to the conclusion that the self model would have to encode not only the
physical possibilities inherent in the body and its interactions in the world". said Mr Holland. "It would also have to include an evaluation of the
significance of the various planned or modelled outcomes".
The researchers will build a robot which will be presented with a succession of increasingly complex environments. To start with, everything around it
will be in fixed positions. Eventually, there will be moving elements and things which will respond to the robots movements. "It will have to model
and predict what they're going to do", Said Mr Holland. "These things will have various values for the robot, some positive some negative".
In theory, the robot will have to construct models of its self and the world around it to survive. In this sense, survival means achieving whatever
mission it has been set by the scientists, which may well be just to accumulate points.
Mr Holland will rely on clues from the robot's visual systems to search for signs of consciousness - in much the same way that the human
consciousness is dependent on our visual systems.
For the rest of the story and other links to go with this project click below: