Facial analysis in process. Photo: Reproduction Instagram Ubu Editora.
Facial analysis in process. Photo: Reproduction Instagram Ubu Editora.

*By Mateus Nunes

Cover of the book "Image Policies: Surveillance and resistance in the dataosphere", by Giselle Beiguelman, published by Ubu Editora. Photo: Disclosure.
Cover of the book “Image Policies: Surveillance and resistance in the dataosphere”, by Giselle Beiguelman, published by Ubu Editora. Photo: Disclosure.

Em The sound and the fury, William Faulkner writes: “I went to the dresser and picked up the watch, still with the face down. I broke the glass on the corner of the furniture and I trimmed the shards in my hand and put them in the ashtray and ripped out the hands and put them in the ashtray too. The ticking did not stop.” The feeling of being equipped with a marker that no longer serves, inefficient in showing a flawed representation of a reality that doesn't stop, is the disquiet felt when glimpsing the world after reading the book. new book by Giselle Beiguelman, Image policies – Surveillance and resistance in the dataosphere (Ubu, 2021). When holding something that ticks and doesn't show the time, we may be carrying a bomb in our hands: cell phones and watches that have the more distant functions of making calls or showing the time. the text of Beiguelman it leaves us breathless, our hearts racing, being careful not to make any sudden movements: we have a grenade of images in our hands.

When exploding, these images become infinite, fragmented and pulverized in what the author precisely calls datasphere. Structurally, these infinite images can be understood as rhizomes or as constellations, in their macro and microcosmic dimensions. Beiguelman brings us, however, a vision beyond the understanding of the structure, but towards the understanding of the dynamics: the images proliferate in a pandemic, in a screaming exponential extreme. Images are contaminated. This contamination of images in the dataosphere, given the reading analogous to the current pandemic, is already seen in psychopathologies - such as anxiety-depressive disorders - and their somatization from digital phenotyping and the repositioning of human subjectivity. No one goes through these pandemics unscathed. In the field of image theory, the duality of the paradigm that Beiguelman presents us goes beyond the ambiguity of the image, which, at the same time as being structural, is also dynamic. The author questions the pathos of the image, both in psychic and pandemic pathology, as well as in the pathos imagery, a concept dear to Western philosophy and aesthetics for millennia.

In astonishing times of fake news and invasion of privacy, it is clear that consumption is intricate to surveillance and scanning practices, which become protagonists in contemporary systems of power. The circulation of data-images is inexhaustible, and this personal information is the product. In this way, the algorithms are fed to better serve as companies that take advantage of them, forcing the user – in the dataosphere, every human is a user – to be bombarded with virus-images. The ruling artificial intelligence simultaneously subverts, inverts and reverses ontological and metaphysical notions about what is true, exemplified by the deepfake, in which case it is confusing to discern whether what we see is real: “The more the discriminator learns to recognize false images, the more the generator learns to deceive it”. Beiguelman's main argument is that, in contemporary society, images and data are infinite, incessant and unstoppable, guiding political and social dynamics.

Within a fog of mirrored noises, like shattered fractals, suspended by an atmosphere infected by fake news in frantic and almost automated routing of WhatsApp, an illusion of equality is created between the “I” and the “other”, drowning in a narcissistic plunge that leads to very serious decisions. According to Beiguelman, Bolsonaro's election was caused by an extreme identification of the then candidate with the habits of his voters: “Throughout the entire electoral campaign, in front of (his own) cameras, the candidate Bolsonaro laughed, became serious, challenged 'the media', prepared the bread with condensed milk for his breakfast, went to the butcher and barbecued. He would show up at the barbershop, pose with his daughter, lounge on the couch and share treats received from anonymous followers. Wearing a sports shirt, shorts, and even a suit and tie, already in the post of president, he doesn't talk to his voters, he expresses them. And, in expressing it, he turns him into a hero, inviting the voter to elect himself.”

Facial analysis in process. Photo: Reproduction Instagram Ubu Editora.
Facial analysis in process. Photo: Reproduction Instagram Ubu Editora.

How, then, to understand a quasi-quantum image, with multiple realities, that escapes the polarization between real and false that we are used to? The resistance forces of images, paradoxically, support their dissemination – the double and ambiguous vector of a new understanding of resistance. Beiguelman establishes a methodological-analytical framework worthy of this new nature – or new artificiality –, declaring an essential obsolescence of positivist ways of thinking, based on data that, even though they were already digital, had a certain analog aspect of data processing interpretation. . So that these images can be processed, they are converted into data that homogenize and standardize angles, framing, behaviors and actions: “What is behind this are the criteria for organizing the data so that they are more quickly 'findable' in searches and the ways in which algorithms contextualize the contents in the specific bubbles to which we belong (something that we do not control and that controls us)”, explains the author.

This organization of data, by trying to convert the unruly matter of the images into numbers, fails. It takes the reductionist path of trying to tame the plural powers of the objects to be analyzed in pure computer data, domesticated in databases that, however much they weave hypercomplex operations, have their input in numbers. As Beiguelman reminds us, “computers cannot see”. It is still necessary to create a platform – still idyllic, utopian – that is capable of processing what one sees when one is “in front of the image”, as Georges Didi-Huberman, a French image theorist, says. It is like trying to read images and write in art history still under the formalist shackles of asphyxiating styles and chronologies, on which the parallelism with Beiguelman's text remains extremely pertinent: “visual contents are mapped by the words that describe them and by the recognition of some patterns, such as lines, densities and shapes”. We insist on continuing to live and feed a world that does not represent reality, that ignores the inherent complexities and inequalities caused. The author, based on her own work in the field of image theory, expands the multidimensional horizons of methodological approaches and hopefully presents us with a necessarily delusional epistemology.

Whenever one opts for a repeated adaptation regresses – in this case, the attempt to convert complex images into quantitative data – chaos ensues and the system breaks down: the world in its current situations, both material and immaterial, is unsustainable. As complex as data processing systems and devices are, they still handle images as if they were objective, which they clearly are not. Evidence of this are the phenomena - or bugs, if we opt for an optimism – of racism and misogyny evidenced by algorithms on social networks. These cases, as correctly analyzed by Beiguelman, are investigated by researchers dedicated to the analysis of AI fairness, a “justice of the artificial intelligence”, which seeks an “impartiality” of these algorithms. According to 2018 statistics from Google and Facebook about their own AI teams, only 10% and 15% are women, respectively. This reflects that, although there is an attempt to inputs impartial, there are always ethical tendencies that “escape” to the way these artificial intelligences think: that of the white man. As Alan Jones pointed out: “Algorithms are not racist – we are”.

Print from a search on Microsoft's Bing exemplifies racism in search engines. The print shows a search for "babies" and shows only white children. Photo: BBC reproduction (2017).
Print from a search on Microsoft's Bing exemplifies racism in search engines. The print shows a search for “babies” and shows only white children. Photo: BBC reproduction (2017).

The images are moved with the intention of colliding, as the German art historian Aby Warburg did, resulting in an enormous discharge of energy, as in a particle collider. They are not just symbolic representations, but are objects in themselves: the image is not the representation of something, it is something in itself. It is reiterated that, just as there was an epistemological revolution in Western thought in landmarks that dictated the pace of society, such as industrial and computer models, it is necessary to write a new manual on how to navigate the world of the dataosphere. 

Em Image policies – Surveillance and resistance in the dataosphere, Giselle Beiguelman makes a literality worthy of Mary Shelley, writing a “contemporary Prometheus”, and the forcefulness that Gilles Deleuze materialized in the time-image e the motion-image. In a society where treaties no longer work, Beiguelman writes an image atlas for contemporaneity.

Read excerpt from the book, click here.

*Matthew Nunes is a doctoral student in Art History at the University of Lisbon, with a period at USP, and an architect and urban planner at the Federal University of Pará.


Leave a comment

Please write a comment
Please write your name