Cognitive Bias and Automation

Pascal Bornet (one of the Intelligent Automation gurus and co-author of the book”Intelligent Automation — welcome to the world of Hyperautomation”) in a recent post on LinkedIn (Here are your 188 known Cognitive Biases!), reopened the topic of the relationship between the complexity of cognitive thinking and the attempt at simplification that derives from the adoption of tools such as Decision Intelligence but more generally of Artificial Intelligence.

This figure summarizes the 188 known and scientifically catalogued biases (BIAS) that could influence the results of a decision-making process.

Although traditional culture gives the concept of PREJUDICE a negative value, cognitive prejudices are all fascinating and I am sure that, by delving into some of these, we will find common behaviors in the world of work.
For example:

Occam's razor — (Occam's razor) fascinating name to explain how often the human mind tells us to choose the simplest possible solution to a problem.

Raise your hand if you have never seen choices based on this prejudice; others lie.

Bike Shedding — (the bicycle shed) describes the tendency to dedicate part of our time, disproportionately, to humble and trivial issues, managing important issues superficially.

Raise your hand if you haven't attended surprisingly endless meetings to make trivial decisions; others lie.

What does all this teach us and what is its relationship with the Cognitive Revolution and Intelligent Automation?

First of all, the first interesting consideration is that the human brain is complex and fascinating but it is equally true that, when a decision has to be made, 188 prejudices risk leading us astray.

On the other hand, thinking that everything can be solved with artificial intelligence algorithms would mean placing too much trust in mathematical models that simulate only part of the complexity of the cognitive process with advantages and disadvantages, for example, deriving from simplifications.

But then how can all this affect Intelligent Automation?

If one of the most centered definitions of the same is that it “automates activities by imitating the four main abilities of workers such as execution, language, vision and thought/learning”, it is instinctive to think that it is practically impossible to recreate with algorithms the difficulty capable of circumventing decisions influenced even only by the known biases of the human brain.

But here lies precisely the fifth essence of the INTELLIGENT component of automation.

Expecting machines to fully imitate the complexity of human reasoning is ambitious, but thinking about automating those activities that do not require, among others, the analysis of all 188 “prejudices” is not at all ambitious, on the contrary, it is a prerequisite for letting people focus on tasks in which the critical spirit of cognitive thinking can create value.

After all, the definition is not limited to the concept of thought/learning, but there are also other elements to imitate such as execution (task automation), language (language recognition, character recognition and BOT) rather than vision (image recognition, augmented reality).

This is Smart Automation!

‍