I'm a pragmatic dualist myself. Evidence from parapsychology has convinced me that 'mind' does things that can't be reduced to physics as we know it, or reasonable extensions of physics (no promissory materialism here please). So the pragmatic conclusion is to investigate mind on its own terms as well as investigate 'brain,' rather than sitting around for progress in brain science to explain (away) mind. Telepathy, for example, is a fact: Sometimes a human mind can communicate information to another mind when there is no conceivable way of doing this physically. Very, very, interesting. My theory of mind must take this into account.

So I have no strong belief one way or the other as to whether a computer will ever develop real consciousness, but I expect to have a wonderful time communicating with smarter and smarter computers, and great fun puzzling whether they are really conscious or not. I think it would be delightful if consciousness developed above some level of complexity, but that wouldn't necessarily be the same 'mind stuff' that human consciousness is composed of.

I see a lot of discussion on this, particularly on the mailing list of the Journal of Consciousness Studies. The main reservation I have about it is that most of the folks doing it are so arrogantly sure of themselves (which means ignoring everything that doesn't conveniently fit in), and they're not having much fun! I love my ideas and theories, but as I've gotten a little more mature, I’ve learned to remember that they are just that – ideas and theories, not necessarily reality.

I've just had a wonderful over-the-telephone coaching session from one of the world’s best meditation teachers, quite responsive to my individual needs. And all run by computer program. This will be one of the biggest breakthroughs in meditation teaching in the century, but I don't think the computer is conscious!

Charles T. Tart. Professor emeritus, department of psychology, University of California, Davis, and professor at the Institute of Transpersonal Psychology, Palo Alto, USA.

My impression on this is that it will be, or might be, possible to have an intellectual machine some day which behaves like our human, in a very limited sense or field.

Shigeki Sugiyama. Softopia Japan, Keio University, Tokyo, Japan.

Will robots ever dream? If we were to take a human brain and replace every synaptic connection between nerve cells with an electronic connection, we could replace a biological brain with silicon. We still wouldn't know how the electronic 'brain' works; many of the mechanisms of perception, memory, etc. have been worked out, but the parts of the brain that make humans human – the motivational and emotional parts – have structures that are buried deep in our evolutionary past and are almost completely unknown.

If this silicon brain were successful in simulating a biological brain, it would certainly sleep, dream, and experience. There is no reason why the content of its dreams should be any different from human dreams. But in a practical sense, there would be no reason to attempt to build such a brain. Human tools are designed to do things that humans do poorly, and computers are no exception. They are good at calculation, for example, but poor at common sense. We already have simple 'androids' participating in many aspects of our lives. We call them traffic signals, bank teller machines, etc. They are designed to do specific tasks to make our lives easier, not to be like us.

Bruce Bridgeman. Professor of psychology and psychobiology, University of California, Santa Cruz, USA

Androids will not dream of electric sheep – because there will not be any androids. The human being is such an ungainly heap of design flaws that no one will waste the time and money needed to make an artificial one. We will, however, fashion devices bright enough to merit the description 'intelligent'. Such devices will have desires (why else would they do anything) and will need to be able to entertain possible ways of satisfying these desires (or else they wouldn't be very intelligent, would they?) – and isn’t this precisely the stuff that dreams are made of? What, then will these utterly inhuman intelligences dream of? Let us hope that it is something as mundane as electric sheep, for they might instead dream of ways to rid themselves of the pesky heaps of design flaws that were their designers. Intelligence thrives on wildness: it does not take well to enslavement.

Pete Makdik. Philosophy-neuroscience-psychology program. Washington University, St Louis, USA.

Although the arguments against computer consciousness based on the Godel theorem are not conclusive, there are serious consequences which arise if one makes the strong AI assumption that human thought is computational. This assumption, coupled with the Godel theorem, leads to the conclusion that there are things in the world which are, in principle, unknowable, something which is in contradiction to one of the basic assumptions of science, namely, that the world is rationally ordered and this order can be known by human reason.

Burton Voorhees. Professor of mathematics, Center for Natural and Human Science, University of California, Berkeley, USA.

I don’t rule out the possibility of machine intelligence at some time in the future, but I believe we need a radically new paradigm for its introduction. In my opinion, none of the current available technologies are going to deliver, and I include artificial neural networks in this assessment. I agree with John Searle that, essentially, computational techniques just deliver syntax and not semantics. The problem of how consciousness arises in humans has yet to be solved, and until we het a handle on that problem, I don’t think we have a clue as to how it can be achieved in machines.

David W. Salt. Head of division of computing, University of Derby, England.

The way we perceive information is so complex and intermingled with other factors which affect a person’s life that mimicking a human brain is really difficult, if not impossible. Thought perception really differentiates the way human psychology works. Things close to nature are more susceptible to such changes in thought perception. To simulate thinking process in a machine, we have to be close to natural processes and a deeper understanding of thinking processes are required. Two people, fed with different info, take it differently than two machines. They act according to their environment. Such behavior has to be simulated in a way that machines can perform it. Behavioral psychology also plays a major role in decision making. In a nut shell, understanding of intelligence is restricted by the fact that intelligence is required to know itself. It is a really goofy vicious circle and there is no way, at present, to get out of it.