Observing Algorithmic Decision-Making: From the Perspective of Sociocybernetics

This report aims to explore a sociological framework for understanding algorithmic decision making and algorithmic decision support, which is becoming increasingly prevalent today.

First, we ask what the interpretive framework for AI is (how AI is interpreted by people, how social systems observe AI, second-order observations), independent of what AI itself is (first-order observations).

Next, we will focus on the point where “final judgment by humans” is the breaking point in the acceptance of decision-making by AI. Then we sort out what is being backgrounded in the series of interpretive schemes concerning AI. For example, who (or what) makes decisions, what is called decision-making, and how is the distinction drawn between AI decision-making and “human” decision-making, and so on.

As a result of the spread of algorithmic decision-making, it is expected that the pattern of social interpretive frameworks of “human,” “self,” “subject,” and “decision-making” will change (evolution of meaning), and that will require sociology to abandon the representation of “human” as a theoretical starting point.

In conclusion, we point out that in cases where we can find a feedback loop between the subject making decisions and the object being decided, the key to understanding this issue is how to position the algorithm within that loop.