Throughout the intellectual development of mankind, the question of how the brain works and how we think has been of immense importance.
From Baruch Spinoza and René Descartes all the way to Karl Marx and Charles Darwin, the question of what makes human beings “conscious” has been inescapable. Yet it was Alan Turing who, by asking “Can machines think?”, changed the course of intellectual debate around this issue and brought about the idea that what seems most random and incomprehensible about the brain may very well be explained, if not recreated, mechanically. As of today, with the rise of Artificial Intelligence (AI), we stand at what is perhaps the most intense and, to some extent, chaotic point in the history of this discourse, where ideas clash and uncertainties grow simultaneously with exponential technological progress.
Can machines “think”?
What does it mean to “think”? Humans have speculated about the nature of thought for centuries, but there is very little disagreement today about what thinking really means.
It is, in layman’s terms, the outcome of a series of processes taking place inside the brain. However, the most revolutionary argument on this matter was presented by Alan Turing in his 1950 seminal paper, “Computing Machinery and Intelligence”, where he proposed what is now known as the Turing Test. The technical details of this work fall out of the scope of this piece, but in a nutshell, the Turing Test was meant to demonstrate that a machine, if carefully designed, can exhibit behaviours that are indistinguishable from those of a real human being. Turing’s work may have been arguably too abstract for his time, but with the rise of AI and chatbots, its great relevance is more than ever appreciated, and his main point has been realised. The question then becomes: if machines can, at least in principle, imitate thinking or behave similarly to human beings, what differentiates us from them? What differentiates our biological brain, forged through centuries of evolution, from the mechanical brains that were once made of copper and wire and now only seem to exist virtually? Thus, the concept of “consciousness” came to be the centre of this debate.
So, what is “consciousness”? To most distinguished scientists and scholars, the term “consciousness” seems ill-defined, and this ambiguity is indeed the root cause of many disagreements in ongoing debates and writings. It is at this point in the history of post-modern science and technology that we witness a rift between scholars. There are those loyal to Turing’s computational views who are taking on a path of rejecting consciousness and suggesting that it simply is a deeper level of thinking, and there are those who continue to believe that something at the core of human beings differentiates them from computers.
Computational reductionism
The computational view has, for the most part, dominated the current paradigm of this discourse, as it also happens to lie at the heart of technological innovations in the 21st century. The original argument was simple: that human behaviour and whatever human beings do is a series of precise algorithmic steps taken in accordance with some biological or psychological rules of inference. But soon it became clear that not only was there no clear biological or psychological argument for logical inference, but that human beings are not constantly following rules of some sort.
Humans are not simply static creatures following rules of nature, but we dynamically interact with what we see, hear and feel. With this view, a whole new movement of scientists, primarily computer scientists and cognitive psychologists, such as Frank Rosenblatt, David Rumelhart— who were later known as connectionists— as well as David Marr, put forward a set of ideas that concentrated on dynamic interaction with the environment and the importance of learning rather than inferring.
The product of decades of work in this direction has brought about, in addition to the field of computational neuroscience and modern machine learning, a significant part of AI, which nowadays is mistakenly referred to as “the AI”. It is important to note that this school of thought recognises the inherent differences between humans and computers, but argues that some form of computation must be the key to the functioning of the brain and that consciousness should be explainable in those terms. Hence, this school does not explicitly reject the idea of computers potentially becoming as conscious as human beings. Further, given that the main building blocks of the more recent theories in this field are neural networks that were in part inspired by the brain, their arguments seem to offer a solid basis for integrating the physical, psychological and biological aspects of the brain and human behaviour.
The rival view
The rival view, primarily spearheaded by renowned mathematicians, logicians and philosophers such as Roger Penrose and Douglas Hofstadter, rejects computational reductionism from a fundamental point of view. The common argument that is often raised is from the perspective of mathematical logic.
Around the same time as when Turing, Alonzo Church and John von Neumann had just begun formulating the mathematical foundations of informatics and computer science, Kurt Gödel, arguably the most important logician in modern history, proved the well-known Incompleteness Theorems. These state that in any formal system, there will always be propositions that cannot be proven using the rules of the system. In other words, it is never possible to verify every proposition within a system itself; there will always be true statements that we cannot prove. In terms of computers, this essentially states that all computers, no matter how powerful, are always limited in scope and that there will always be problems that a computer cannot solve.
The argument of this school of thought is based on the fact that human beings, especially in mathematics, have a record of going beyond the scope and limits of computation, namely the development of mathematics itself or the solution to the uncomputability of the tiling problem. Therefore, human thought— or consciousness, if you prefer— must go beyond computation. They instead view the brain as a physical system operating by laws and theories that have not yet been understood. For instance, Penrose’s argument is that consciousness is non-deterministic and therefore arises from quantum processes inside the brain that are not yet fully understood by physicists.
AI and consciousness
Where do the breakthroughs in AI stand in this debate? At first glance, they do confirm the importance of learning from data and actively interacting with the environment, although the success of robotics and more advanced computer vision is yet to be tested. However, there are clear limitations of machine learning systems in performing reasoning and logical inference, which do not entirely align with human intelligence.
Chatbots and other highly intelligent machine learning systems depend on significantly complex neural networks, many times bigger than that of the human brain, and yet they depend on enormous amounts of statistical data. And still, there are tasks that AI systems are either just as good in performing as humans or not good at all. This suggests that the computational view, while on the right track, may not have the ultimate answer. And likewise, the rival view, while arguing from a rigorous mathematical perspective, fails to take into account the dynamic and evolving structure of human behaviour.
Perhaps there is a clear need for a new alternative. Instead of clashing the two classic schools on what consciousness is and whether it exists, we should attempt to merge the two ideas and acknowledge that both point to important aspects of the human mind. Theories such as the Relativistic Brain Theory put forward by Miguel Nicolelis and Ronald Cicurel, while incomplete, give us clues about possible formulations where the brain is not viewed as just a computer—which mathematically seems unlikely. This new perspective, where computational performance of the brain at different levels is observed and explained, is formed by viewing the brain as a hybrid system.
The lessons in the modern history of science point towards successes that were brought about through the contrarian fusion of ideas, from Albert Einstein’s relativity to the development of quantum mechanics and even the development of the theory of computation itself. Science has moved forward thanks to merging often opposing views. With this in mind, it would seem that the real challenge of uncovering the true mysteries of the brain lies ahead for the next generation of researchers and scholars.
