By Louis Fourie
JOHANNESBURG - Some time ago I was watching an ant erratically moving around on the beach, wondering what it was doing. It reminded me of the ant parable of Herbert Simon, the American political scientist, cognitive psychologist and a Nobel laureate in economics.
The ant parable
According to Simon in his book “The sciences of the artificial” an ant moves laboriously across a wind-shaped beach to his goal. While moving ahead, he constantly changes direction to climb a dune, circumnavigate a pebble, or exchange information with a fellow ant. It is obvious that he does not follow a straight line or the shortest route to his destination. The reason is that although the ant has an underlying sense of direction, his movements are irregular because he cannot foresee all the obstacles and must continually adapt his course to the problems he encounters. His horizon is very close, and he deals with each obstacle individually without giving much thought to future obstacles.
For the human observer the path of the ant looks erratic and incredibly complex. But the complexity of its behaviour and path is not due to a complexity in the ant itself, but rather a reflection of the complexity of the surface of the beach or the environment in which it finds itself. To understand the ant’s behaviour we really need to understand the contribution of the beach or the context of the ant.
Similarly, the complexity of human behaviour over time is mostly a reflection of the complexity of our environment.
In the case of a human being, a problem-solving environment is often described as an extensive maze of possibilities that is selectively searched and reduced to manageable proportions.
The problem is, however, that human beings have a limited capacity to store chunks of information in the short-term memory (often referred to as “cognitive strain”) and, therefore, find it difficult to execute an efficient strategy unless the stimuli is greatly slowed down or they could use external memory aids.
Although humans have virtually unlimited long-term memory for the storage of a large variety of information, it takes time to transfer items from the limited short-term memory to the large-scale long-term memory. Due to the memory limitations of human beings it could thus be hypothesised that human goal-directed behaviour mostly reflects the environment in which it takes place, as in the case of the ant.
Based on the thinking of Herbert Simon, Rodney Brooks, a roboticist and emeritus professor of robotics at the Massachusetts Institute of Technology, contended many years later in his famous “Intelligence without representation” lecture that animals are not so smart because they do complicated mathematics in their heads, but because they rely on the world as its own representation. An easy analogy is the typical monthly groceries shopping at the supermarket. You can either buy according to a pre-made list or just walk through the aisles and pick up items that prompts your brain when you see them.
The arguments of Simon and Brooks agrees with that of the much earlier Sigmund Freud, the Viennese psychoanalyst, who stated that man is not at all rational, but rather driven by instincts and desires, just like any other animal.
Dependence on AI and search engines
The above limiting traits of human beings are the reason why people are becoming increasingly dependent on artificial intelligence (AI), algorithms, machine learning, and even search engines to solve problems.
Unfortunately, AI and web search engines have fundamental limitations.
Both are excellent in answering “what is” questions that portrays the current state of affairs based on historical facts and statistical calculations, such as: How long would a trip by car from Cape Town to George take?
Search engines, for instance, merely do massive lookups of an indexed database according to an algorithm.
Despite tremendous progress, computers have difficulty in answering “what if?” questions and applying imagination or common-sense reasoning.
Some scientists try to solve this problem by building deep learning networks with numerous layers requiring massive computing power. Unfortunately, the limitations of data science and AI is that they both solely focus on historical data.
Improbable and the Matrix
These limitations are partly the reason why a UK start-up called Improbable, is following a fundamentally new approach and is attempting to build “The Matrix,” a next generation virtual reality (VR) world like in the 1999 American science fiction film “The Matrix”. The virtual reality is not created in some abstract sense, but genuine, living, breathing recreations of the real world that allow people to have totally new experiences. Improbable, for instance, built a virtual replica of the city of Cambridge, with 130,000 virtual inhabitants. It included simulations of the traffic and public-transport networks, utilities, power lines, mobile-phone and internet systems.
Improbable’s ultimate goal is to create totally immersive, persistent virtual worlds, and in doing so, change how we make decisions. Improbable, therefore, released a software package called Spatial OS that enables programmers to simulate immensely complex interactive systems with millions of components. They also teamed up with Google to simulate the world using Augmented Reality.
Now let us get back to the ant and its complex environment, the beach. The Matrix is an ultra-high definition simulation of the real and complex world that entails the modelling of the interaction with hundreds of thousands of people, cars, buildings, animals, objects, the weather and many more. The Matrix makes a different approach to AI possible, where intelligence materialises from the synergistic interaction of simple entities embedded in complex environments. This will allow us to ask “what if” questions of The Matrix, as well as questions that requires imagination and common-sense reasoning. The Matrix will then simulate the questions and provide possible solutions to complex problems.
According to this view, intelligence is not considered as an ability inherent to a living creature, but as a composite of the interaction of the creature with its complex environment.
This advanced simulation of The Matrix may indeed bring many new insights to individuals and organisations, showing that businesses often do not follow the theoretical tenets of expected utility theory, but rather use simple heuristics that may not necessarily be optimal, perfect, or rational, but is nevertheless adequate for reaching an immediate, short-term goal or approximation.
The current Covid-19 crisis and society’s response to it would be a good example. Perhaps The Matrix would have helped us to take totally different decisions – decisions that are more rational, not driven by our fears, instincts, desires, or immediate short term goals, but decisions that lead to actions that maximise our expected utility.
This view rings so true when we see society’s response to the Covid-19 crisis. Do the past few months increase our belief that humans are rational beings who take actions that maximise their expected utility? I do not think that Herbert Simon would have been surprised at all by our response to the pandemic or by many of our daily decisions and actions. Due to our limitations in short-term memory and speed of stimuli in the complex modern world, our erratic and seemingly complicated behaviour does not necessarily reflect human intelligence, but it merely reflects the complexity of our environment while we are navigating daily obstacles in our personal life and in our businesses.
The goal of AI should thus not only be to represent some mythical and uber-rational super human intelligence and reasoning ability, but to understand that like all other animals, also human “intelligence” owes much to the complexity of the world we find ourselves in. We are more driven by our needs, instincts, desires, fears and imagination than we would like to acknowledge.
How improbable it may seem at this time, The Matrix may lead us to a more achievable technology in the long run. It could mean the end to the popular search engines and AI as we know it. Unfortunately, the company Improbable are experiencing some strong headwinds due to some bugs with their Spatial OS. They have lost some clients and are still struggling to make their mark in the world of games, which is the testbed for The Matrix. But when they succeed in the future, our world will be very different.
Hope to see you soon in The Matrix.
Professor Louis C H Fourie is a futurist and technology strategist.