At HERC we publish blog articles covering a wide range of issues that broadly relate to harm, evidence, crime and justice. In keeping with the critical position of HERC, our aim is to highlight all sides of the debate and to facilitate a discussion so that all voices are heard on the issue.
In this blog post, Dr Lee John Curley discusses the widespread fears of AI as involving the loss of ‘our special human capacity of rationality’. Lee John Curley is a lecturer in Psychology at The Open University.
Our fear of Artificial Intelligence once related to the terminator and Skynet, but in a time of economic uncertainty and mistrust over how artificial intelligence uses human data on the internet, new fears are more related to employment and human rights. Some people see the development of AI as a process in which we recklessly hand over our special human capacity of rationality to machines, condemning ourselves to low paid jobs, or even unemployment. In this week’s blog, I explain why psychologists are interested in rationality. I present the fable of Prometheus, the great titan who was punished for passing on his godly skill of rational thought. I highlight the lessons that can be learnt from this story when considering potential implications of artificial intelligence.
Rationality or the ability to integrate information to choose an option with the most utility, is a cognitive ability that may be at the heart of what makes us human: the very meaning of the term Homo sapien even means “wise man”. Rationality has become such as constant in human behaviour that the pillars of society (law, economics and medicine) all assume that decision makers employ rational processes when faced with an option. This blog will delve into how the ancients viewed rationality, how modern cognitive psychologists view the term and how rationality will shape the future.
However, rationality has been studied by more than just cognitive psychologists. Mathematicians, philosophers, social psychologists and psychoanalysts have all studied rationality, each with different viewpoints on rationality and the extent to which humans participate in rational behaviour.
In Ancient Greece, the world was explained in terms of symbolic entities (gods, deities and titans) that represented observable phenomenon. For instance, Gaia represented the earth, Poseidon the seas, and the almighty Zeus was symbolic of the heavens above. Some of these powerful beings, however, represented very human traits. Prometheus (meaning forethought) and Epimetheus (meaning afterthought) represented the rational and non-rational (or intuitive) part of the human mind, respectively. Once these titans fell out of favour with the Olympians, however, their roles of rationality and intuition fell to the gods Apollo and Dionysus. Prometheus was the champion of thinking ahead and choosing the long term right path, despite the negative short term effects for himself. This is evidenced in the story of Prometheus where he steals fire for the ancient humans, against Zeus’s instructions, and is punished until he is freed by Herakles. Despite the negative ramifications for himself, he metaphorically, and literally, ignites rationality, abstract thought and logic into the minds of Homo sapiens; thus simultaneously making humans more like the deities they worshipped, and the gods less special. The creation of the Prometheus myth shows that rationality is a key aspect of humanity, and that the ancient Greeks were aware of the power of rationality.
During the Renaissance, there was a reawakening of rationality, with mathematical (or normative) concepts, such as probabilities, essential to modern mathematical and psychological theories of rationality being invented. With rationality and probability becoming interlinked, humans were viewed to be “Laplacean demons”. In other words, ‘we’ were viewed to be rational beings, who had unlimited cognitive capacity and were not influenced by the limitations of the mind. In association with this development in rationality and mathematics, institutions such as law, medicine and economics were all developing fields and were influenced by the perspective of the time (i.e., to be human was to be rational).
This was the main viewpoint until the cognitive revolution in psychology and the seminal work of Tversky and Kahneman. They conducted a number of experiments in the 70 and 80’s (and even won a Nobel Prize) for highlighting that although rationality should govern our minds when making decisions, that instead, individuals sometimes deviate from rational principles and make decisions based on intuitive cognitive short-cuts called heuristics (Greek for find or discover). Their research showed that humans are flawed and that we can make biased decisions.
This perspective has dominated the majority of the last 50 years of work in the field of decision science. Contemporary decision scientists, however, see intuitive thought and rationality as brothers (similar to the Greek myths surrounding Prometheus and Epimetheus). The dual process model of decision making suggests that two different modes of cognition (system 1 and system 2) governs our decision making. System one is an intuitive mode of cognition with a plethora of heuristics making up the components of said system. System two on the other hand is the rational part of the mind, which may be unique to humans. System two is believed to be more effortful and conscious than the primitive system one mode of cognition. The modern mind-set of rationality is that it is possible to make rational decisions, but that it is difficult and effortful, thus researchers believe that humans much prefer to default to system one.
This flawed perspective of human rationality has led to rationality, the very essence of humanity, becoming synonymous with artificial intelligence and robotics. Normative (mathematical) models of rationality have been shown not to reflect the entirety of human behaviour, whereas artificial intelligence (AI) may be a new frontier to apply these classical models of decision making. Unlike human beings, artificial intelligence can be programmed to accord with rational principles and statistics. Therefore, what classically was seen as something unique to humans, the thing that made ‘us’ special, may in the future become a robotic trait. This mirrors Prometheus’s gift to ancient humans which lead to deities becoming less godlike, and humans becoming more like their creators.
Now computers are powerful enough to win against a human at chess, and it is estimated by researchers that AI will exceed human ability in a number of tasks (e.g., language translation) in the next 10 years. It is even believed that by 2053 AI could replicate the abilities of a surgeon. This speculation suggests that the expansion of artificial intelligence into the realms of rationality may cause humans to become obsolete, with more rational, consistent, and efficient computers replacing biased and flawed humans. This could cause a number of occupations traditionally employed by humans to be performed by complex AI.
Others, such as Peter Fleming, instead argue that AI will cause an increase in poorly paid jobs, as he argues that an important factor in AI being utilised in a profession is, will it be economically viable? Therefore, Fleming suggests that low skilled and low paid jobs will not be replaced. He expands on this point by suggesting that AI that partially automates a job though an app will also reduce the skill required by the employee, thus decreasing the relevant pay required for the service (e.g., Uber driver with app vs. traditional taxi driver that receives training). Furthermore, contrary to contemporary belief, the age of the AI may have a negative effect on human standards of living. Humans, like Prometheus, may suffer the negative consequences of passing on the sacred flames of rationality to an intelligence that ‘we’ created.
In summary, rationality has always been viewed by humans as a god like ability. The story of rationality is the story of humanity, the way we view rationality changes how we view ourselves, and ‘we’ are becoming increasingly closer to mirroring the story of Prometheus and igniting the flame of rationality in non-organic decision makers, and thus decreasing the specialness of humanity. By giving this special ability to AI, we may be condemning ourselves to low paid jobs; or even unemployment. Further bringing to life the story of Prometheus, as the great titan who was punished for passing on to humans his godly skill of rational thought.