On BBC programme “VISIONS OF THE FUTURE”, Michio Kaku, a Japanese-American theoretical physicist, discussed the possibility of robots with human-level intelligence. One of the “humanoid” robots he introduced in the programme, was Honda’s Asimo. This robot can see, listen, speak, and walk. It greets to a visitor, takes him to a caffeteria, brings a coffee to his table and has a bit of chat. If the progress of the technology continues, one day we may encounter robots with human-like conscousness with real emotions, like happiness, sadness and anger. And they may start requesting legal rights equal to humans. So, are you prepared to recognise intelligence, consciousness and emotions in robots and if you do, will you accept them as a member of our society?
What I have to say here is that we don’t know enough about our consciousness, to unequivocally judge the possibility of robots consciousness. Nevertheless I still think robots lack a crucial ingredient of consciousness, which is self-consciousness. They lack it because robots are made for purpose, namely to serve us. Robots with real consciousness would place themselves in the centre of their mind and act on their own accords to maximize THEIR chance of survival. Of course it is not the scientists but the public who decide whether they want such “selfish” robots or not, and they must ask themselves whether it is morally consistent to build and accept new artificial conscious agents, in our society while rejecting potentially concious life being born from us, namely human.
While this is a very trendy topic, and might be an appealing thought to lots of sci-fi fans, we are no closer to creating consciousness in a laboratory than we are to creating life. Indeed the former (the creation of sentience) is probably orders of magnitude more complex than the latter (the creation of life). One might even speculate that there is a larger reason for why Nature is so difficult to penetrate, and so reluctant to give up its secrets – namely that an ethically immature species (hard to argue that we are not still ethically compromised as a species) is not ready for the knowledge and enormous power that goes with it. Additionally, given the evidence that the management of homeostasis is somehow central to the creation of a sentient state (and presumably consciousness was selected evolutionarily because it improved homeostatic management and the maintenance of life processes), it seems very difficult to construct a conscious architecture in a robot that has no real homeostatic processes. The notion that consciousness will be relatively easy to create an artificial life form in a few years or even decades frankly shows no respect or humility in the face of Nature. We know far less about the neural substrates of consciousness in staggeringly complex biological systems than we think we do, and to instantiate this process in something other than a biological brain is something that we are no closer to than interstellar faster-than-light travel. In other words, I think we have much more immediate scientific concerns and I think it will be a long time before your vacuum cleaner says to you “I feel devalued when you turn me off at night.”
Thanks for your comment. I think your argument is valid as far as human consciousness is concerned, but in this blog entry I intended a more basic form of consciousness. While many would agree that consciousness is not unique to human, defining the minimum mechanical requirements for consciousness is far less straight forward. In his paper in Annual Review, John Searle mentioned three elements of consciousness; qualitativeness, subjectivity, and unity. Unfortunately these are notoriously difficult to measure and don’t help us much to determine the mechanical requirements. Alternatively, we can pursue a bottom-up approach, where man-made objects will be tested for their “conscious” behaviours. Management of homeostasis, for example, already exists in some robots which recharge their battery when it is low, but of course it is not enough to say they are conscious. So what’s missing then, or what should be added to them next? That’s the question which I hope I can address in the future in this blog.