Barnes in Common

the magazine of Churches Together in Barnes
September/October 2007

Home
Current issue
Previous issues
Noticeboard
Discussion board
About CTiB

Can a robot be like a human?

John G Taylor (Emeritus Professor of Mathematics, King's College London) with a reply by Ross Collins

To answer let me start from the human brain. There's an enormous programme on brain science being carried out world-wide, using ever more sophisticated brain measuring equipment (PET, fMRI, EEG and MEG machines we now read about constantly). Allied with single cell measurements from monkey brains, this research is leading to an enormous increase in understanding of the human brain. With the vast growth in computing power now becoming available (think of a million computers connected together) we are seriously beginning to replicate the activity of the whole brain. What are the glimpses we are getting of how our brains work (and beginning to be implemented in robots)?

First there is emotion, claimed to be at the basis of 'being human' and the core of many religious faiths: 'God is love'. Emotion is now being shown to reside in various brain components in what is termed the 'limbic' system. Models have been constructed of how predicted reward values are created in our own limbic system inside our brains, using the important brain chemical dopamine (thought as the basis of drug addiction). These models are being used to develop emotional 'appraisals' (and related emotions) of how various aspects of the external world can thwart or supply solutions to highly valued goals.

We also importantly possess cognitive faculties - reasoning, thinking, recalling from memory and speaking. All of these are controlled by our amazingly subtle faculty of attention. This acts as a filter to cut down the complexity of the surrounding world and help concentrate on a single stimulus at a time. Robotic software with a visual system under attention control (guided by reward values of goals) has been created, for example in the GNOSYS EC Project I am co-ordinating at King's College. Attention can also function to help various executive functions to be performed, including reasoning, thinking, remembering and speaking. All of these are under active investigation, and our GNOSYS robot has already a primitive reasoning system up to the level of the chimpanzee, and presently being upgraded to that of a two-year old.

The software systems being created can learn like an infant, with a tabula rasa initial structure, and can thence acquire the abilities of the adolescent and so, we hope, attain adulthood. It is important to have robot development guided by associated rewards/values/ emotions so as to lead to reasonable, non-psychopathic behaviour. I repeat again that the robot would be 'brought up' like an infant - it would not be programmed in any computer fashion to choose good versus evil at the whim of the programmer, but have to make that choice autonomously for itself.

But what about consciousness - the soul - how could that be implemented in a robot? I suggest that our 'inner self' is created by some sort of 'wake-up call' - an 'ownership' precursor signal about the contents of consciousness you are about to experience when you attend to an object. This wake-up call arises as a copy of the attention movement signal, and allows scientific predictions about how it occurs in the brain. Science can thereby probe consciousness itself and is presently doing so by brain science.

Would there be spirituality in such robots? With complex enough brains, there is no doubt they would start to speculate and some might even postulate a God (or an Intelligent Designer) to explain away the puzzles of the world. Others would try to understand as much as they can rationally. Not much different then from humans!

Further reading: The Mind: A User's Manual, John Taylor, pub John Wiley 2006

'Can a robot be like a human?'

a reply from the Reverend Ross Collins

It was a bit of what I call a 'Barnes moment' when I talked to John Taylor at a cocktail party and discovered he was a world-renowned pioneer in the development of artificial intelligence and robotics. We had a compelling conversation about robots being theoretically capable of being as human as you or I, and I was delighted when John agreed to write an article for us and for me to reply from a theological standpoint. (See above)

Can a robot be human? Clearly, there are human qualities that a robot can perform. I am sure a robot can reason, learn, think, in a human way (and probably better than human beings) - the computer which took on the chess champion, Gary Kasparov, is a good example. More controversially, I can concede too the possibility of a robot having human emotions. It should be possible to replicate neural pathways and processes such that a machine could respond with anger, joy, etc.

However, at this point, I part company with robotic humanity. Many of the key experiences of humanity would not be part of a robot's.

John catches the problem when he talks about robots being developed to behave in 'non-psychopathic' ways. The paradox is - we would not want a robot to do evil, yet programming a robot so it cannot commit evil would remove a key human capacity - that of having to choose between good and evil. A robot would not be a moral being.

Could a robot fail? I don't mean 'crash' like a computer. I mean 'fail' in the sense of not achieving what was expected, letting someone down, feeling guilty, etc. Could a robot be capable of originality and creativity, reach the musical heights of Mozart, the artistic brilliance of Caravaggio? I know the vast majority of us never attain such originality, but the presence of the exceptional within the broader community of humanity is, to my mind, a defining characteristic of who we are.

Finally, can a robot be spiritual? It is one thing to say it could postulate an Intelligent Designer (though would that not be its programmer rather than God?), ie, have an intellectual belief in God's existence. It is another for it to have faith - to put its trust in God in the self-giving, risk-taking way characteristic of authentic human experience.

Back to the top