Integration of information as a function of consciousness ​

To build a model, you need to combine or integrate the received information. This is how it happens: information for the model is selected, then analyzed, combined according to certain principles, and finally synthesized. At the output after integration, we get a model that can be used to make changes in activities. For example, changing the way to attract customers to the company. Or even choosing a different role altogether: for instance, after describing roles in a project, realizing that the role of a product architect is not being fulfilled, and stepping into that role.

According to modern theories of consciousness, generative models - descriptions of the world based on which we act - are always generated by the brain. Part of this process happens automatically, unconsciously. For example, when someone tells us "Ivan will come soon," we turn towards the door and expect Ivan to walk through it shortly. In our model, the appearance of the object "Ivan" in the room through the object "door" is anticipated.

The most complex models are usually consciously constructed, in S2 mode, often collectively. For example, the layout of a passenger airplane is created by a team of engineers. It includes numerous interconnected objects and takes longer to create than the model of "Ivan's arrival."

Since models always exist in one form or another, it is desirable for these models to be good. That is, they should have good predictive power - when applying models, agents should encounter less uncertainty and make fewer mistakes. This is the essence of the Free Energy Principle the brain tries to biologically implement. We can assist by learning modeling and keeping good models in memory.