The case for greyscale in the age of AI
‘The art of meditation is a way of getting in touch with reality, and the reason for it is that most civilized people are out of touch with reality because they confuse the world as it is, with the world as they think about it, talk about it and describe it. For on the one hand there is the real world and on the other there is a whole system of symbols about that world which we have in our minds. These are very very useful symbols — all civilization depends on them — but like all good things they have their disadvantages, and the principal disadvantage of symbols is that we confuse them with reality, just as we confuse money with actual wealth.’
This quote by the zen philosopher Alan Watts, reflects on a fundamental truth that the certainty of symbols and knowledge is only a representation of reality. But this approximation comes from capturing the essence of true reality, which is uncertainty, in the form of knowledge. It is at that uncomfortable edge of uncertainty that all human wisdom and knowledge has arisen.
Uncertainty, prediction and the human brain
According to a new theory of cognitive philosophy called Predictive Processing, reality, as we experience it, is built from our own predictions. All human experience is part phantom — the product of deep-set predictions, and, we can no more experience the world ‘prediction- and expectation-free’, than we could surf without a wave [A]. This theory goes by other names as well — hierarchical predictive coding and active inference. New sensory signals are used to refine and correct the process of informed guessing (the attempts at prediction) already taking place in the brain. Predictions and prediction errors are thus increasingly recognized as the core currency of the human brain, and it is in their shifting balances that all human experience takes shape. This contemporary picture of the predictive brain can find roots in the 19th century German polymath — Hermann von Helmholtz, who suggested that we perceive the world only thanks to a kind of unconscious reasoning or inference, in which the brain is asking itself, ‘Given everything I know, how must the world be for me to be receiving the pattern of signals currently present?’ Try to listen to a familiar song over a poor radio signal and the words and rhythms can still be surprisingly clear. This is not so with a brand new song. This is your brain’s predictive modeling at work. If the world does not send strong counter evidence, those predictions sculpt experience, making the song sound clearer to you.
However, the perceiving brain is never passively responding to the world. Instead, it is actively trying to hallucinate the world by checking the hallucinations against the evidence coming in via the senses. Thus, the brain is constantly painting a picture, and the role of the sensory information is mostly to nudge the brushstrokes when they fail to match up with the incoming evidence. For, the way our bodily states feel to us reflects a complex mixture of what our brains predict and what the current bodily signals suggest. This means that we can, at times, change how we feel by changing what we (consciously or unconsciously) predict. Our experience of reality thus rests on this uncertainty and the capacity of our brains to predict. The possibility to exist in this greyscale is very crucial for us. Much of what we think, feel, and do, arises from this carefully controlled hallucination.
David Marr, a towering figure in the world of computer vision, developed a model that goes like this: visual processing starts by detecting basic ingredients in some incoming signal. From there, layered processing slowly builds towards a more complex understanding in a feed forward manner. This has been extended and used in the neuroscience domain over the past many decades. However, since the backward propagation has been missing in the original model, this has not been carried out into the neuroscience world. The links from the deep brain down towards the sensory organs has been put aside. However, the number of neuronal connections carrying signals backward in this way is estimated to exceed the number of connections carrying signals forward by as much as 4:1. It is a puzzle that we’ve evolved to maintain such a huge network of neuronal connections feeding from the brain to the sense organs. Given this, the brain requires an extensive energy outlay (up to 20% of our bodily energy consumption) to sustain its activity to maintain a model that issues moment-by-moment predictions.
When inner guessing completely rules the roost, we are just hallucinating, full stop. But when it is appropriately sensitive to sensory stimulations — via prediction error signals — the guessing is controlled, and the world becomes known to the mind. Guessing is what defines the greyscale!
Certainty and discomfort
At the University of California, Berkeley, psychologists began giving people the cat-dog test, where they started with the image of a cat and iteratively morphed it into a dog. Participants were asked to state when they perceived the animal in the image change. In these tests, it was seen that the participants refused to surrender the safe harbor of their first answer until the morphing was nearly complete. Else Frenkel-Brunswik, a researcher in the study wrote, ‘each time, the participants showed a preference to escape into whatever seems definite.’ What she found as a key signature of a closed mind was the intolerance of uncertainty.
Those who shun the indefinite tend to see the world in shades of black and white, us versus them, ignoring the grey. They are prone to jump to answers and are distressed by chaos and surprise. Their cognitive map is narrowed down to rigidly defined tracks. In contrast, people who operate on the other side of the scale are more likely to be curious, flexible thinkers who revel in the complex problems and in new experiences from living abroad to trying a new delicacy. They may even be in better charge of their minds and studies indicate that such people have more grey matter in their brains.
While on the one hand, it is human tendency to seek to eliminate uncertainty often, evolutionarily, most transformative changes have occurred when uncertainty has been embraced. Think of it this way, if you are not willing to admit your own limitations of knowledge, you run the risk of being surrounded by yes-people, and never learning. Imagine being stuck in a place where you never learn, for the rest of your life. Humans share with animals, about half a dozen core instincts such as to flee, fight, coordinate and reproduce. Over millennia, these moments of instinctual slowing evolved into the human capacity to direct our fate by consciously working through the unknown, something that the psychologist Jerome Bruner calls, detachment of commitment. In the words of Aristotle, ‘It is the mark of an educated mind to be able to entertain a thought without necessarily accepting it.’ By committing foolishly to a viewpoint before considering all aspects, one does not give oneself the chance to become educated.
‘A pause is not a time when nothing happens,’ says Stanford neuroscientist Vinod Menon. In an insightful experiment, his team eavesdropped on the thought patterns of participants to baroque symphonies. To their astonishment, the brains were most active during the morsels of silence between movements. When the certainty of the next step ceased, their brains were pushed to a cognitive cliff’s edge — the place where new learning happened. When we stop running away from uncertainty, we can tame the fear of not-knowing that keeps us at a toxic remove from a multifaceted, evolving understanding of reality and that stunts our capacity for imagination and creativity.
Routine experts versus adaptive experts: why being comfortable with uncertainty is important today?
Were the most respected and valued thinkers of the past, such as of the renaissance or any of the other golden eras of any of the past human civilizations, people who were always certain or, was frailty considered an important aspect of being fluid enough to receive and transmit the truth? Is a thinker who hesitates in crisis considered foolhardy or heroic? Are some heights of excellence best left only to the resolutely sure? Who will we look to for wisdom, in the next free fall of civilization? Who are the great thinkers of our day?
‘Intuition is nothing more and nothing less than recognition,’ concluded Herbert Simon, a 1978 Nobel laureate in economics. To see wherefrom this notion arose, we need to go back to 1939, when a Belgian freighter set sail to Argentina, with some of Europe’s chess masters. On board also was Adriaan de Groot, a budding psychologist. His studies on board and subsequently, helped define two levels of cognition. One that was evolutionarily ingrained in us — the likes of approaching a stranger cautiously (cognitive system I). The other was self-taught from accrued experience at an individual level (cognitive system II). The ancient adage experto crede — trust in one who has had experience lends itself to this notion. However, the heirs to Groot have chanced upon more complex aspects of the brain. The mind is far more than a machine. Superior thinking under fire demands more than push-button shortcuts that may have been acquired by cognitive system II. And the very know-how that makes us an expert, is all too often our downfall. This is often called the Einstellung Effect — after people master one particular method of solving a problem, they become blind to better routes to solve similar problems. The certainty of our own knowledge can make us unaware of new learnings.
Modern experiments have shown that how we set the question context in our brain, heavily influences how our neurons attempt at solving the problem. Before heading there, we must visit the practical experiments that Karl Duncker, a Gestalt psychologist visiting the United States pursued. His findings let him conclude that tapping one’s memory for past solutions has a certain heuristic value, but had little to do with thinking. Efficiency in problem solving, he found has a costly side-effect. Tweak the rules of the game that someone has spent years mastering, and they are prone to perform at a loss. In a modern day version of Duncker’s famous candle experiment, Tamsin German tested at which age this fixed thinking may come into play. In kindergarteners, he saw that they were able to approach the problem with a ‘what can the object do?’ rather than ‘what is the object made to do?’ attitude, which helped them solve the problem often more effectively than adults! Their innate willingness to be unsure helps them solve the problem faster. Routine experts fall into the creeping intuition bias, says psychologist Anders Ericsson, while adaptive experts become unsure under duress, thereby welcoming the space of uncertainty as a launchpad for assessing a challenging situation. What looks like hesitation, is in fact a controlled effort to reconcile new with old knowledge, to flesh out the early impressions, and to reveal not needless complications but the crucial complexity that is already there.
So, going back to the question that this section began with, the most proficient artisans, says artist David Esterly, inhabit a world permeated by error. By seeking to extend rather than just apply their knowledge, they ward off the arrested development of routine expertise.
The long haul: being human
While uncertainty is very crucial, constant uncertainty can lead to stress. In groundbreaking work by neuroscientists Jacqueline Gottlieb at Columbia University and Robb Rutledge at Yale, they independently found that the brain adapts by a process known as focused arousal, whereby the brain deals with uncertainty by trying to be the most stressed by the most unpredictable situation, thereby reducing the constant background stress. In clever experiments designed by them, they saw that the best players were those who adapted the most to uncertainty. By best, I imply in terms of lessened average stress and better game outcomes. Humans who adapt to uncertainty are on average less stressed out. At the University of Washington at St. Louis, Ethan Bromberg-Martin conducted a series of cognitive experiments on monkeys to conclude that information — any information — is so critical to our survival that a range of brain regions sounds a steady drumbeat favoring its pursuit. Evolution has endowed us with an innate sense that in almost every situation, we can do better if we first know more. We are drawn to uncertainty since that is where additional information can be extracted. This is how babies explore: strategically, but without knowing the highly significant role of play. He concludes that we are exactly the sort of creature who ought to have systems to figure out now just the value of things that we know, but to value the exploration of the unknown, even if that seems useless at the moment. While psychologists have long held the belief that learning is a process of associations built on repetition, there is growing evidence that it is the uncertain and unexpected, that is the starting point of expanding our knowledge. As stated in my previous post, entropy is linked to new information via the unexpectedness of a piece of data, so this new conclusion does not seem awry.
Alan Turing once remarked, ‘Processes that are learnt do not produce a hundred percent certainty of result; if they did, they could not be unlearnt.’ In the age of AI, such intelligence becomes our routine experts in the black and white landscape of facts and knowledge. To be human then, is to be the adaptive expert who flourishes in the greyscale landscape of wisdom of the unknown, who accepts an opinion different from theirs, who has empathy towards another and who embraces the uncertainty as an evolutionary friend.
(Disclaimer: all views expressed in this are my own and do not represent those of any organization that I may be affiliated with)
Bibliography:
A. ‘The experience machine: how our minds predict and shape reality’, Andy Clark, Pantheon Books, 2023
B. ‘Uncertain: the wisdom and wonder of being unsure,’ Maggie Jackson, Prometheus Books, 2023
