For a few decades the relationship of our bodies with digital data has become increasingly intimate and, at the same time, problematic. In an extremely dataified world, we have begun to feel anxieties related to the inability to maintain coherence between the scales of digital abstractions and our specific somatic experience : likes and followers.they accumulate but we feel more and more alone; the rhythms of work and the economy are accelerating, but this does not imply any “progress” in our working lives but, rather, an increase in competition and precariousness; The Big Data market has turned us into merchandise, but we do not receive any kind of retribution for the economic value generated by our data, nor do we have the possibility of democratically choosing for what purposes it is used; social networks allow us to communicate directly and on a global scale, but their use encloses us in custom algorithmic boxes ” filter bubbles “; We have more information than ever in human history, but misinformation reigns as fake news proliferates .
It would be, however, naive to propose a return to the somatic body without further ado and deny digital mediations and abstractions. In this session, coordinated by the researcher Alejandra López Gabrielidis (CNSC / Tecnopolitica ), we will explore, first of all, to what extent this tension or digital anguish can be the basis from which to think about an expansion of the spectrum of our corporeality, using the theory of individuation of Gilbert Simondon as a framework and tool for reflection.
Can data be thought of as an exosomatic body ? How to articulate the somatic body and the data body? In what way can the reality scales that each of them open be made compatible?
Second, we will investigate other possible modes of coupling between bodies, data, and algorithms. One of the most ubiquitous and influential current forms of appropriation of our bodies of data is the technology of machine learning , a specific way of building models from learning programs that manages to classify new data to fit an implicit pattern in a preset set of training data. These are programs that “learn” from examples (positive and negative) to recognize similar examples.
These model building techniques are fairly elementary and far from exhausting the possibilities that programming provides, but they are immensely effective. It is necessary to consider the reasons for such effectiveness in the field of human action both to review the very conception of the human , and to project possible futures with better couplings between humans and programs.. It is usually assumed that automations reside in the delegation of human decisions to algorithmic processes, but human decisions themselves are not exempt from automatisms related to prejudices, social and cultural differences, power structures, etc. What happens, on many occasions, is that these pre-existing human automatisms are revealed and favored by these technologies.
There are, however, other types of couplings between humans and algorithms, such as the one that occurs in (meta) programming processes , that is, the construction of programs that build programs – that allow the realization of sociotechnical systems with high degrees of indeterminacy, fostering new and promising forms of cooperative human-machine creativity.
- Date : May 26, 2021
- Place : Canòdrom -Ateneu d’Innovació Digital i Democràtica
- Event registration here
Materials of interest
- Video session available here
Photo: The Vector