Artificial Intelligence and the Evolving Theory of Homo Economicus
– Aditi Sharma
The principle of rationality and homo economicus has always been the basis behind the development of economic principles. Neoclassical Economists believe that homo economicus is the first and most basic agent of optical decision making. The assumption behind homo economicus is that an ideal human being behaves rationally and makes optimal decisions for himself considering all the possible actions, preferences, and consequences of its action. This assumption is based on a mathematical connection between choices and values. Behavioral Economists, on the other hand, argue that humans possess bounded rationality centered on simple heuristics. Heuristics psychology is based upon the fact that human makes some instantaneous decisions which are not entirely rational. These decisions are what make them alive.
The advent of technology and the invention of artificial intelligence has exposed a new aspect of homo economicus. Artificial Intelligence has brought new arguments relating to rationality. It has not only helped humans in reaching a homo economicus-like ability but has grown to an entirely new economic agent machina economicus, or computational rationality. It has substituted human decision-makers with machine intelligence. This intelligence considers an additional anticipation of negative consequence. Its impact on economic models, which explains the social, economic, and political outcomes of decisions, made it work better.
The critics of machina economicus argue this concept in three respects. Firstly, technology lacks elements like cultural knowledge, reasoning, or common-sense, which play an essential role in making rational decisions. A common example is a game of poker. Playing poker requires not just knowledge of rules but also soft skills, like to analyze your opponent’s psychology, his behavior, and talents like bluffing. But, to our surprise, a new software technology Libratus defeated poker professionals over more than 100,000 hands of poker games, indicating that a machine can possess not just hard but also soft skills. Secondly, the inventors have neglected that some decisions cannot be solved optimally. For instance, solutions to trolley problems, or what is commonly referred to as an ethical dilemma. If a self-driving car has to choose between colliding with a group of five people and one person, what will it choose? There is no optimal solution to such a problem. These ethical dilemmas are long debatable subjects; nevertheless, technology cannot be trusted with the decision making of such problems. The problem does not end here. Another question relates to the attribution of liability, if such mistakes are made by artificial intelligence. Who would be responsible for the outcome of such decisions? Lastly, the possibility of fooling technology by misleading the input in the software. For example, a sign indicating veering into the left lane made by sticking stickers on the road to confuse the intelligence of a self-driving car is, in fact, detected by it. This prank causes a huge accident. A student from the University of California, Berkeley, researching on this says, “These problems are more concerning than idiosyncratic quirks in a not-quite-perfect technology”. The reason for considering this possibility is that studies have demonstrated that fooling the software or manipulating it by adding or removing some errors is easy and prone. These technologies are fundamentally brittle. The research is still incomplete regarding this matter.
Artificial Intelligence has both promising as well as challenging effects in economics. Humans have always had a fear of losing control over decision making and evolving machina economicus is one such step towards the fear. The basis of economics vests upon individualism, as supported by Adam Smith. It means that humans have always been self-reliant. They have always chosen to be responsible for their own decisions and well-being. However, the shift towards machina economicus challenges our quirk of self-reliance. Machina economicus is evidence which reveals that humans and machines are inseparable, and humans have themselves designed a method that potentially interrupts their choices by machines and its intelligence to antithesis bounded rationality. When the attempts are successful, the choices made by humans will be based upon suggestions made by artificial intelligence.
The emergence of a new type of economy due to machina economicus is expected to result in more efficiency and optimality. But complete chase of a ‘rational mind’, even through technology, is still a long way. This aspect still requires more research and study by experts. Though we are finding ways to overcome all possible difficulties in our way to reach optimal decision-making through machina economicus, this is not enough. There is a need for technological-infrastructural changes and properly structured guidelines to help in fulfilling the aim of machina economicus to achieve rationality and fairness in decision-making. We must have theories that can describe its behavior. There must be some standards that should be accepted by society, which can help the builders to make rational decisions while dealing with problems like fooling the software.
It is said that with great power comes great responsibility. Researches have shown that artificial intelligence can become a menace and a promise to society, both at the same time. What would it turn out will depend upon the developers of that technology, the agency relationship, and the economics of scale. We will, soon develop the technologies that would take decisions on our behalf, benefiting the consumers and granting them stretch to enjoy other choices and actions. But it cannot lead us to become those agents that would make us the best economic decision-makers. The power would always lie in the hands of those machines. We will develop an entirely new type of economy, economica machina. Our role would be to avoid the pitfalls that may arise in the way, and we are waiting for that stretch.
This article is authored by Aditi Sharma. She is a student of law at Maharashtra National Law University, Nagpur.