Can Computers do TOK?

If some people are to be believed the world depicted by such sci-fi blockbusters as 2001: A Space Odyssey, Blade Runner, AI, I Robot, Ex Machina … is one we will soon (or already do) inhabit. It is a world in which machines will increasingly exhibit human traits to the point where it will not be possible to decide whether one is interacting with a person or a robot. Alan Turing famously raised the question as to whether computers can think in his 1950 paper, ‘Computing Machinery and Intelligence’. Turing quickly had to change the original question to ‘Are there imaginable digital computers which would do well in the imitation game?’ as it became clear to him that even the computers of his day were already performing operations which were akin to ‘thinking’. Turing’s own answer was that computers would be able to successfully pass his test by the year 2000. This is a highly charged and volatile debate and one which often polarizes opinions but it seems that AI experts are for the moment and by and large siding against Turing. So, if computers can not yet think, can they do TOK?

Before answering our main question let us review what AI experts say about the possibility of ‘thinking computers’. The main centre of human thinking is of course the brain. This extraordinarily complex organ is, however, so difficult to describe that metaphors have been the standard way of doing so. The most common simile has been to see the brain as a kind of computer, or as a network of computers. This seems to have the merit of neatly capturing some of the most obvious operations of the brain, but also to account for the most puzzling aspect of the brain’s operation which is the mysterious interaction between its material side (electrochemical impulses) and its immaterial side (thoughts). The material part can then be seen as the equivalent of a computer’s hardware and the immaterial as the software. This model is, however, increasingly being abandoned as AI experts claim that if the brain is a computer at the very least it is an biological one and therefore its operations are more organic than purely mechanistic. This is particularly true in the case of creative innovations which are difficult to account for solely as the product of algorithmic processes.  For many AI experts these are two of the most profound challenges to the idea of ‘thinking machines’. Firstly, the human brain is much more than just a computer; secondly, they do not see how modern technology could possibly replicate something which evolution has fine-tuned over millions of years.

But TOK is not just thinking: it is also a very peculiar kind of thinking. To be able to do it at all or to be really good at it requires more than just computational abilities, it requires a self. The Theory of Knowledge makes clear that knowledge in all its aspects, from understanding its very nature to fathoming how it is produced, acquired and applied, is more than the manipulation of mere data. The apprehension of knowledge is a highly complex process which involves much more than just the purely logical. The interplay between all Ways of Knowing is unique for every human being, and one could argue that the context in which knowledge is produced is also highly specific and may therefore be unrepeatable. But data is not only information: it also acquires meaning, and that meaning is contingent on a myriad of factors some which will be connected to what is commonly described as emotional intelligence. This is something which, some would argue, it is difficult to see machines ever possessing.

However, as the saying goes, only the fool says ‘never say never’; it may be that emotionally intelligent machines are looming on the horizon; on the other hand, it may a while before students can ask their computers to write their TOK essay! You do need a brain to do TOK, but maybe you also need a heart and a soul.