Experiment results: how to make a robot manager learn more about what a human manager does?

Posted by

In 2015, I conducted a series of experiments using the same online manager tools, which I slightly modified programmatically. However, their function was not changed, which boiled down to measuring the organizational quantities of the primary ones – goal and task.

The first of the series of experiments was conducted again among students of the management specialty at the University of Economics in Katowice. The participants of the experiment were given an elaborate case study of a company intending to change its current office to another one with the characteristics set by the experimenter. Using the managerial tools Goaler and Tasker, the experiment participants had to plan this venture considering two criteria, each with two alternative states:

  • taking additional corporate social responsibility measures or not taking them (criterion A),
  • implementing the venture within 2 weeks or 3 months (criterion B).

The experiment involved 16 managers (students) in charge of three-person social groups.

We performed a linguistic analysis of the value of the magnitude of the recorded managerial activities aimed at assessing the intertextual consistency of the instruction that the participants received in the experiment with the effects they achieved at the end of the experiment.

The main conclusions derived from such analysis were as follows:

  • the more thought processes the study participants had to perform, the more accurate were their descriptions of the goal or task placed in the managerial tools, meaning that more of the measured quantities of their managerial activities were recorded by the managerial tool, and also that the accuracy of the values of the specific measured quantities was higher,
  • in a situation where the undertaking to be carried out was communicated in a low-order manner, the reconstruction of detailed tasks and objectives to be carried out was quite a complex mental process and consumes a relatively large amount of managers’ time,
  • the more details were included in the command to execute the venture, the less details the participants included in the final result (the final versions of the goals and tasks) while creating content themselves that was not present in the command.

It can be added that the accuracy of the description of the goal or task in the project increased as the number of thought processes the participants had to perform increased. It follows that if the more we want to teach managerial activities to a robotic manager, the more we need to induce thought processes in human managers to more accurately express what they are thinking with the content written in managerial tools. Let me remind you: managerial tools are also measurement tools!

You can read more on this subject here:

Flak O., Alnajjar J.: Linguistic Analysis of Managers’ Behaviour Aimed at Replacing Human Managers with Robots