SOPHISTICTED AI MODELS
This is a true story and I knew some of the people in it; but I have changed the details to avoid offence. I do know that the person's parents are dead and that she had no children.
Georgina was a brilliant student. She got a first-
Yes, Georgina's models could predict the future movement of financial markets alright; but the one major movement they did not predict was the global economic recession that took place on the eleventh of September 2001. That was the day when terrorist attacks on the Twin Towers of the World Trade Center in New York caused 2,977 fatalities. Sadly, one of these fatalities was Georgina. She now lies in a grave, well away from New York, but next to her parents.
*******
The essential lesson that we learn from this story is that all systems, whether financial, biological, political or whatever; have environments. In this context, an environment consists of those forces outside a system, of which it has limited or no knowledge, and over which it has little or no control. So the environment of a system can seriously disturb it, or even destroy it, with little or no warning.
Perhaps the most well known example of modelling these days is the modelling of our changing climate. Like Georgia's modelling, this has been very successful in following the systems that it is designed to model and indicating what potentially influencing actions can be taken within it. But just as Georgia's model could not predict terrorist actions, so climate models cannot predict, for example, major volcanic eruptions which could blacken the sky for years and make global warming only history. (Please note that I am not denying the need for action on global warming).
The essential point about models is that they are based on human knowledge, experience and understanding: as it is at the time when they are made. Yes, the use of AI can build learning into the model that can redirect and expand its scope; but here again, it is human decisions that have built into the model the range of its awareness and form of its learning, based on past experience.
So once again:
The only 'law' of management is that management is limited in what it may achieve
and that management is not always aware what these limitations are.
