Metrinomics as a service provider for employee and customer experience measurements, was always blessed with data that was at the core of human interaction. Due to the confidentiality of this internal data, it was not possible to conduct classic scientific research on it.
We had to have our own laboratory for this. Fortunately, one of the strengths of AI is that it can be measured against reality. Therefore, it has always been possible to prove that a process works, even if it is not yet known why and how.
Thus, in 30 years of analyzing human behavioral data using AI techniques, a body of rules has emerged. These rules stand out because they are confirmed repeatedly, so that they seem to be universally valid.
They also confirm each other independently of languages and cultures. Whether Europeans, Asians, Africans - what they differ in is little, what they have in common is much.
This is how the view of an artificial intelligence on the human species could be summarized. But what does this common ground consist of?
Insight 1: Humans may not be predictable, but they are readily groupable. That is, everyone is an individual in the way he behaves, but there are always groups of others who behave very similarly.
Insight 2: Humans are 85% social beings and 15% individuals. This means that 85% of our behavior can be explained with a few simple rules, and for the remaining 15% much complex logic would be necessary to find rules in it.
Insight 3: Humans strive for consistency. This is an inescapable basic behavior of all. It means that a person's attitude or behavior on one issue can be used to infer other issues. It also means that under 'normal' circumstances, it is very difficult to lie and deceive consistently. It costs a lot of energy and is therefore readily avoided. Nevertheless, lying is part of everyday life insofar as it reduces social stress, which also requires a lot of energy.
Insight 4: Man can focus well and differentiate situationally, but cannot store this. Only a faint memory image remains of each situation, which fades more and more with time.
Insight 5: Man learns as long as he/she lives. Even when situations fade, the moment leaves a trace in the learning process of life experience. This trace is often detectable even if the image, the memory of how it came to be, has long since been extinguished. This is one of the reasons for 'irrationality' that we observe in the behavior of others.
Insight 6: Man is a victim of his/her perception. That is, he perceives only a section from the 'reality' which corresponds to his just active sense organs. This perception is sent in addition still by the consistency filter and colored by interests. The result is considered as 'truth' and it is assumed with pleasure that the perception of others cannot be different than the own.
It cannot be said for what there are no words. It cannot be imagined for which there are no images. It is an eternal endeavor of man to harmonize his own subjectivity with that of others in the social structure. Therein lies an essential driving force for social achievements, such as art and science, and thus for the individual a source of satisfaction and a sense of achievement.
Insight 7: Man strives for uniformity. Changes in attitudes and behavior occur according to rules of swarm behavior. It is more attractive to join others than to seek one's own way.
Insight 8: Man needs the distance to others. Equally important as harmony is demarcation. Distancing forms abilities that make the individual useful for the community and strengthen it. This is true not only for the individual, but also for the group.
Insight 9: Wellbeing is an individual goal for all and sundry. It is a striving for a balance of conditions that is the prerequisite for one's own prosperity.
Insight 10: Prosperity has a place, and this place has an attraction. Places have a prosperity value, which from the subjective point of view describes the conditions it offers to the individual. Individuals aspire to places with a high prosperity value and strive to increase the prosperity value in the place where they are.
Insight 11: In the (subjective) evaluation of what is experienced, scales are used to grade between 'very bad' to 'very good '. The question is, how can the information be graded so that as many subtleties as possible are transmitted. For this purpose there are 5, 6, 7, 10 and 11 scales in use.
The fact is, behind each of these scales there is actually a 3-point scale. In the end, it gives the clearest picture in the analysis. Everything that is more complicated requires abstraction and sometimes succeeds better, sometimes worse.
Insight 12: When using AI to analyze social interaction, generalizing helps more than specializing. Using AI to personalize directly is nonsense. Generalization contains 85% of the information.
Generalization also has the advantage that the privacy issue can be greatly simplified. It is not the individual that is of interest, but the group to which the individual can be assigned.
Generalizing means placing the integrating in the foreground, the personalizing based on this focuses on the separating. Just as Homo Sapiens likes to see himself in his social structure.
Insight 13: Perfection does not exist. It is not even desired. From the perspective of a learning AI and in working with one, the most important performance indicators are the speed and efficiency of learning, of adaptation. Adaptation only works if there are deviations, i.e. errors, to deal with. To get an AI to the optimum, the logic of things leads to wishing for optimal errors.
This already has elements of nonsense, but it points to one thing: the data must be organized in such a way that errors are clearly recognizable as such and they can be processed by the AI in their data structure. Deficiencies at this point rob the AI of an important part of its power.
From this point of view, perfection is an extreme deficiency. If everything runs flawlessly, there is nothing left to learn.
However, because this is unrealistic, it is ultimately an unattainable goal that actually stands for the path: fewer and fewer errors. Experience has shown that it is incredibly difficult to raise satisfaction (as the absence of errors) in social relationships from 90 to 100%. Much harder than from 40 to 50%. So that means, what is good, can't grow better any more? This also has elements of nonsense. It suggests much more that in such a case the goal dissolves, loses its meaning. There is nothing more to learn.
Looking on all the history of trials to eliminate 'errors' in social (employee and customer relations), it shows that emotionally, it is the process of improvement that captivates, not perfection. Perfection is the end of development. Perfection is considered cold. What is static and remains static, is dead. AI says: the goal of perfection belongs on the dung heap of history. Or is it already there?