FROM CYBERNETICS TO THE METAVERSE: A GENEALOGY OF ALGORITHMIC CHARACTERISTICS, TRANSPARENCIES, AND OPACITIES

Abstract

When we think about digital technologies, the Internet, and its influence in the different spheres of social life, it is usual to point at the informatic algorithms as the central pillars of these effects. Several discursive and practical deployments expressing hopes and fears have been deployed around these figures, mainly because of their potential for transformation, but also because of possible biases in their design, development, and application. However, computer algorithms, as a digital technical object, present their own sociotechnical concretization processes that have led them to be one way and not another. The objective of this article is to establish a genealogy of the evolution of computer algorithms, setting five sociohistorical periods to observe and analyze the changes in their conceptualization, production, and characteristics, as well as the central aspects around their transparency and opacity. Thus, it starts from the Macy Conferences of 1946 —where cybernetic criteria were established regarding what a computer algorithm should be like— until reaching the current situation where several types of algorithms, many of them imbued with machine learning and deep learning processes, are central participants in the exercises of governmentality and algorithmic personalization.