José Luiz Braga 338 other words, they offer some service. But there is always room for criticism and the search for corrective revisions. Algorithmic selections made by the majority or by the distribution of previous receptions, on average, certainly work. Since they are “automatic”, they encourage their simple accep- tance and reduce reflective and active effort, exempting people from seeking other options or courses of action (which may be better or more suited to singular situations). One likely conse- quence is the reinforcement of behaviors – which will be reiterated even when the conditions of the context and circumstance change significantly. The risk, already present in the processes of cultural standardization of behavior, stimulating conservative tendencies, is amplified with digital processes, given the combined force of the impersonality of the algorithm and the false impression of objectivity resulting from an algorithmic “automatism”. The challenge here is how to avoid automatic selec- tions that reinforce our recent preferences, which can become a barrier to diversification and a better-directed historicity of actions and discourses. We assume that social diversification is important to perceive and address emerging problems and seek new solutions. c) Limitations to creative learning The interactional digital algorithm can learn – in the sense that it constantly adjusts to variations in social processes in the original dimension of its attention. But it is not sensitive to unpredicted dimensions. Or if not, it is only capable of “learning” new dimensions (not envisioned in its original plan- ning) to the extent that these new dimensions emerge and socially establish themselves outside the triggering of the algorithm and, still in the social environment, develop a correlation with the programmed approaches. All of this is dependent on achieving a sufficient proportion of reach for the algorithm to trace and include these dimensions. In other words: the algo- rithm only “learns” what has already been creatively learned in the social context by its users – who would only then provide this renewed data.
RkJQdWJsaXNoZXIy MjEzNzYz