Interactional digital algorithm 337 suit the interests and objectives of the developers, does not provide certainty as to their best relevance for the user participants. These individuals “offer” their data without even having clarity about what is being collected and without being able to gauge the bias of the selections made by the algorithm. On the other hand, the feedback to the user creates the impression that they are being individualized – but, in reality, they are just taken as a point on a statistical curve. Leonard Mlodinow (2009, p. 131), in a book on proba- bilistic calculations and randomness, critically demonstrates the inadequacy of applying predictions viable only for a large number (population or representative samples) to individual or small group situations. Bringing this perception to our object, we can see that the participants (the “public”) are apprehended and served by the dimensions selected and the variations con- structed, when in fact they are being much more framed in the data set than expressed by them. The challenge corresponding to this angle of the issue is how to ensure that the program is pertinently attuned to the social dynamics and characteristics (dimensions and variations) that are relevant to the algorithm’s field of action and to its users. b) Previous choices overempha ize What an interactional digital algorithm captures, in its specific sphere of action, are recent preferences and choices made by its users. The algorithm’s response to these preferenc- es – organizing their diversity and specific variations – is to provide processes, positions, and products based on the frequency distribution of choices already made. This obviously reinforces the maintenance and routinization of these preferences. There is a tendency to discourage new variations, experimentation, and new learnings about the world. This risk can be considered a reducer of social plasticity in the constant construction of its history. In the cultural environment, it is true that selections are made, and decisions are taken without necessarily being the optimal ones – only because, on the one hand, they serve dominant forces and, on the other, they “work” for the problem and objectives at hand – in
RkJQdWJsaXNoZXIy MjEzNzYz