‘In high school I came across the writings and books of the logician and mathematician Raymond Smullyan, and found them to be a wonderful introduction to how logic can be used to represent reasoning processes. This was my first introduction to artificial intelligence (AI), although I was not aware of the term at the time. The work by the Cognitive Scientist Douglas Hofstadter also inspired me – especially his wonderful popular science book Gödel, Escher, Bach: An Eternal Golden Braid.’
These are the words of Prof Tommie Meyer, who is a Professor of Computer Science at the University of Cape Town (UCT), Co-Director of the Centre for Artificial Intelligence Research (CAIR), and leader of the Knowledge Representation and Reasoning Group at UCT.
Prof Meyer was born in Randburg and his pre-PhD work was done at the Rand Afrikaans University (now the University of Johannesburg). He obtained his PhD from the University of South Africa and when he moved out of Johannesburg, it was to settle in Pretoria. He also lived in Pittsburgh (USA) and Wollongong (Australia) for shorter periods before moving to Sydney in 2003. In 2007 he moved back to Pretoria, and in 2016 to Cape Town. He is married to Prof Louise Leenen, who also has a PhD in Computer Science (Artificial Intelligence). They have two sons, Thomas and André.
Prior to his current roles, Prof Meyer held positions at the CSIR in Pretoria, and National ICT Australia (now Data61), the University of New South Wales in Australia, the University of Pretoria, and the University of South Africa. He is recognised internationally as an expert in Knowledge Representation and Reasoning and is one of only three South African computer scientists to have obtained an A-rating from the South African National Research Foundation.
Prof Meyer talks about the people who influenced his career most: ‘My PhD supervisors Willem Labuschagne and Johannes Heidema both influenced me greatly. Sadly, they have both passed away. However, the person who had the biggest influence on me was the late Professor of Computer Science and Engineering, Norman Foo. He was a colleague, but a friend first and foremost. He was my mentor, both personally and professionally. While living in Wollongong and Sydney I was fortunate enough to be able to spend time with him. Whenever I have a difficult decision to make, I tend to sit back and ask myself what Norman would have done.’
At a time when AI has become a general topic of conversation among experts as well as those who are less scientifically versed in this field, Prof Meyer provides some very clear insights: ‘The field of Knowledge Representation and Reasoning (KRR), as the name suggests, is concerned with two aspects: firstly, the way in which we structure and represent our knowledge about the world, and secondly, the way in which we use these structures to reason about the world and make decisions.’
He continues: ‘The representation part is important, since it allows machines to explain and justify the decisions they make to humans. That ability helps to curb any potential biases creeping into the decisions made by machines: if we as humans know how a machine came to its decision, we can decide whether that decision is morally acceptable to us. And the reasoning part is important because if the decisions are based on faulty reasoning, the explanations and justifications will also be faulty. This means that these decisions will not be acceptable to humans. To summarise, the only way in which we can ensure that humans will trust the decisions made by machines is if machines can explain their decisions to us, and those explanations are acceptable to us.’
Importantly, he also explains a common confusion of terminology: ‘Perhaps the biggest misunderstanding when it comes to artificial intelligence is the conflation of intelligence with consciousness. There is a tendency to assume that as machines become more intelligent, it will inevitably lead to consciousness. It is, of course, possible for an entity to be both intelligent and conscious. But it is certainly also possible to be conscious, but not intelligent! For our purposes, the relevant case is the one where machines are highly intelligent, but not conscious. This is where we are heading towards, and where AI already affects the lives of many people.’
He adds: ‘The best way to think of AI is as yet another useful technology like calculators, dishwashers and smartphones. Think of AI tools as labour-saving devices with the potential of making our lives easier, provided we use them appropriately.’
Indeed, AI has already become embedded in our daily lives almost without us even being aware of it: ‘While AI technologies can potentially be enormously beneficial, there is also the possibility, as with all technologies, of AI doing active harm. Perhaps the biggest danger is that the ugly side of being human – our biases – can be carried over to AI tools. We have already seen examples of this, with AI tools exhibiting discriminatory tendencies. This is largely a consequence of the data-driven approach underlying many AI tools in which patterns (and biases) in existing data are picked up and learnt. While learning from existing data is important and useful, it is important not to embed our human biases into these tools. And it is here where KRR can play an important role. Researchers are looking at ways of combining machine learning techniques with the techniques used in KRR to ensure that harmful biases in data can be recognised and avoided.’
On being a scientist – and more
Prof Meyer also shares some thoughts on being a scientist: ‘I am a computer scientist and I spend most of my professional life on aspects of that discipline. But being a scientist means I don’t need to restrict myself to that. I am free to pursue any of my intellectual interests. As a result, I have also done some work in philosophy and even mathematical biology. This really adds to my quality of life. But perhaps more importantly, this type of inter-disciplinary research frequently drives advances in science. It allows scientists to connect the dots, so to speak.’
He believes that institutions such as NITheCS play a crucially important role in a country such as South Africa ‘where it is necessary to pool our intellectual resources and skills in order to achieve impact. NITheCS allows students and researchers from universities across the country to collaborate and share knowledge.’
On a completely different note, Prof Meyer shares this anecdote: ‘I made a brief, uncredited appearance in the Academy Award-winning movie Searching for Sugar Man. I appear in the movie for perhaps three seconds, when a photograph of me is flashed onto the screen. The photograph was taken at one of the Voëlvry concerts held in the late eighties as part of the alternative Afrikaans rock movement.’