'There is no future, only Pattern Recognition!'

In my first encounter with the popular AI contraption, 'Chatgpt', almost a year back, I mustered some courage to ask of her (It behaves feminine, I’m told!), 'what is your instrument of knowledge?' Pat came back a broad reply whose essence revolved around, 'Pattern Recognition!'

The idea of 'pattern recognition' as a core faculty of human mind was first put forward by Nobel Laureate Herbert Simon. Through his experiments, Simon began to understand both human and artificial intelligence as being dependent on an ability to process information and store it, which leads to pattern recognition. Intuition had largely not been understood up until this point, but this realisation caused Simon to hypothesize that intuition was really people using information from situations and scenarios they had previously encountered to respond to new situations. Since they remembered what had succeeded and what hadn't, intuition could lead to making good decisions, even if the decision making processes happened at a more unconscious level that people couldn't explain.??

I was kind of familiar (well before the present AI context of 'Pattern Recognition') with the phrase, in a more nuanced context of 'horrors of the historicity of the past while eliding the reality', having already read an 'off genre' masterpiece by William Gibson titled 'Pattern Recognition'! The message that Gibson carried through the book has been well summarised under the rubric, 'There is no future, only Pattern Recognition!' Gibson’s narrative closely aligns with Francis Fukuyama’s seminal work, ‘End of History’.? Fukuyama saw the world to have reached an epoch whereafter the future would just be a repetition of the past. It essentially meant that there is no future, future is already foretold in the patterns of the past! But the 9/11 event, post Fukuyama’s conjecture, turned the thesis around! It begs a question whether AI is just a past digger, nay, a pattern digger, of a more sophisticated caliber and a significantly better one?

Pattern Recognition also has a very wide reference in Psychology, recognised fundamentally as a neuro cognitive process that involves matching information from a stimulus with information stored/retrieved in/from memory. Human Pattern Recognition can be considered as a typical perception process which depends on knowledge and experience people already have. It’s a survival machinery in perpetual motion. And here lies the important distinction, it is built into and integral to our perception process. Whether AI is endowed with any perception process, we haven’t a clue on that as yet.

There are many theories, e.g., Templates/Prototypes/Features……….and the ones anchored upon statistical gymnastics of the Bayesian kind. In perception, one of the most famous schools is that of Gestalt Psychology (Holism) associated with the adage, ’the whole is greater than sum of its parts’. ?Gestalt principles, for example, are laws of human perception that explain how we simplify complex images by grouping objects and recognising patterns. ?It centres around the idea our perception of the world is based on how we put together and interpret individual elements as one whole rather than considering each element separately. That’s how kids make sense of the ‘whole’ and therefrom, ‘parts’ and vice versa!

?I am not too sure if organisations can address their ‘complex problems’ without ‘Gestalt AI’/Modular Artificial Intelligence, as against the typical mainstream neural network based models.? The decision making behaviours on complex problems are not atomistic (and static) but demand attention to an overall configural nature of the problem. In perception, the Gestalt concept of the organisation allows us to address the dynamic interrelations among the elements constituting a ‘stimulus ensemble’.

?I do not contemplate use of AI in resolving/mitigating routine confusions of organisational life - any of the available non-AI system (including the Legacy system suitably coupled to some advanced technologies) is good enough. The question is whether AI can assist in ‘high level decision making’ beyond and above these routine confusions. ?A CEO’s daily routine, for example,? is a maze of ‘complex choice behaviours’ besieged with paradoxes. ?An example runs thus - some CEOs prefer a sure gain to a risky but possibly larger gain, a process known as Risk Aversion. Yet some others (or, may be the same), prefer a larger loss than a smaller sure loss, known as Risk Seeking. We thus prefer a sure gain over an uncertain gain of equal or larger Expected Value (Risk Aversion). Conversely, we prefer an uncertain loss over a sure loss of equal or smaller Expected Value (Risk Seeking). These instances obviously hint at people’s bias in Risk Aversion in gain and Risk Seeking in loss situations, independent of Expected Value (EV has no meaningful effect in these scenarios). ??What it implies is that Expected Value criterion is obfuscated by a ‘Psychological Meaning’, a perceptive faculty in decision making! These two tendencies can lead to ‘inconsistent’ choice behaviours in decision making. ?These are ‘mathematically’ inconsistent ?but highly consistent otherwise. This inconsistency cannot be explained without a Gestalt view of the perceptive process and cannot be remedied without Gestalt intervention. ?

In organisational situations, things that bring probabilistic returns/gains are usually repeated in the hope that they will be eventually rewarded. That’s why we resort to repetition of the same commercial ad to sell a merchandise and you send multiple applications to get a job! Note that it is not ‘once’ only! It is the perception of overall outcome from many trials and not one trial (whose result is usually negative) that sustains the activity. In the loss situation, the reverse is true. The risky alternative becomes less attractive or more aversive, when it will be repeated many times. People would certainly be more willing to take the risk if they don’t have to take it repeatedly. ?Far fewer people would probably buy insurance if they are ‘covered’ to the risk only once (and the insurer is willing to take the risk thinking that his client will suffer only once at max or not suffer at all)! It is the perception of over-all long term outcome, and not the outcome of one individual event, that motivates a Risk-averse reaction in a loss situation.

There is another ‘perception’, integral to pattern recognition, is ?the ‘perception of consistency’ (and perception of constancy) which is seemingly a less (?) benign armament but with a long range. We are persistently ‘biased’ in such a way that the world we perceive is less variable and more regular than it is. ?For example, we can cover a building far away just by a finger, which means the image of the finger formed in the retina is bigger than that of the building, but if we move the finger away, we can see the building. It is much bigger anyway. The consistency/constancy of perception refers to perception that remains the same when the condition of perception changes in a certain range. And the range varies from person to person….from intellect to intellect. ?

In many situations and of necessity – because of paucity of data, time pressure, or working memory limitations -we have to infer on the characteristics of environment from the sample data. Therefore, we find relationships between actions and outcomes, or between pairs of variables, where little or no correlation exists. With respect to correlations and contingencies, we tend to see them stronger than they really are. Perceiving variance as smaller and correlation as stronger than they are (actually they mean the ?same thing) allows us to see the world as more predictable (less variable) and consistent. ?Seeing the world as more consistent has some obvious implications: if variances are small and correlations are strong, confidence in one’s ability to make accurate predictions and good choices increases. A Gestalt speculation is in play here.

Our tendency to perceive meaningful connections between things that do ?have and do not have any, depends upon how critical are we with our thinking process. ?