Å®ÉúСÊÓƵ

Big money is being spent on initiatives like the European Union’s Human Brain Project. Will people hoping to learn about consciousness be disappointed?
Absolutely. From what I hear, some of that project’s neuroscientists are disappointed because it isn’t nearly strong enough in asking cognitive questions. It is asking the basic, materialistic questions – such as which cells connect with what, or which chemicals are diffusing – but these basic questions aren’t the only important ones.

So are we much closer to grasping consciousness than when you started work on it, four decades ago?
Not very. I think the fundamental problems aren’t just scientific – knowing what’s going on in the brain when we’re conscious and so forth – but philosophical questions, and in particular about the phenomenon of consciousness. This concerns the so-called hard problem of how conscious experience emerges from matter, and why we experience, say, the redness of red or feel pain. It isn’t just that we’re not sure what scientific questions to ask; it’s that we don’t know what questions to ask because we don’t know what we’re talking about.

What about brain imaging studies. Don’t they help?
Most work going on in brain imaging is of no scientific value. It looks for correlations between behaviour or experience and activity in the brain, but it’s rarely guided by theoretical questions. Maybe in 100 years, it will fit into a neuroscientific theory. Right now I regard it as natural history rather than science, in the sense that Darwin turned natural history into theoretical biology in On the Origin of Species. Most imaging work is at that natural-history stage.

So where have we seen progress?
One area is in understanding functional consciousness, such as decision-making. And we understand more about how systems in the brain cooperate and integrate to make conscious or unconscious decisions.

Would you ever expect to see the sort of human-like artificial consciousness of Ex Machina or Chappie?
I don’t think it’s impossible in principle, but if it does happen, it’s going to be very far in the future. Apart from requiring very powerful computers – and soon some computers will match our brain’s processing power – we have to understand the brain’s information-processing well enough to make a system that can do it. We simply don’t.

How does this limit research?
In terms of AI, it means you have to pose strictly specified problems before coming up with systems tailored to solve them. In many cases you can solve such problems far better than any human can, and your solution will have amazing applicability – but only in a very, very narrow sense.

How do you get round this narrowness?
The brain and consciousness, like climate change, is one of the few areas where people realise that different disciplines have things to contribute. This must be encouraged. The problem comes with publishing results – if you don’t keep within recognised research boundaries, what you’re doing is not recognised.

But you bucked that trend?
I had a very strange educational background and this is partly why I’ve been able to make a reasonably unique contribution. My first degree was in medical sciences, then I studied philosophy, and social and cognitive psychology, and picked up AI. These days people like me might not get a job.

Tickets to the discussion event at the Edinburgh International Science Festival on 7 April are