Bruno Moreschi’s work involves deconstructing systems and decoding social procedures and practices in the fields of the arts, museums, visual culture, and technology. The artist seeks to offer other readings or even to explore the internal structures of organizations, beyond their official discourses, as he did in 2018 with his project for the 33rd Bienal de São Paulo – Affective Affinities.

Entitled Another 33rd Bienal de São Paulo, Moreschi’s project created analyses of images from the Bienal by Artificial Intelligence; audioguide tracks with commentary from the exhibition’s personnel; enlargements of Bienal texts; unconventional filming of the assembly, and recordings of the public’s reactions.

Another 33rd Bienal de São Paulo, 2018

I think the project had a certain discretion that fascinates me to this day. When Gabriel [Pérez-Barreiro, curator of the 33rd Bienal] invited me, I first thought it would be impossible for me to build something in dialogue with the power (sometimes authoritarian?) of Niemeyer’s architecture of the Pavilion. That’s why I loved the fact that we would have a website and that the project is really the content created there. Also that, as soon as the Bienal was over, it would somehow become part of the institution’s archive, in other words, it would have survived modernist architecture. Some Bienal visitors were amazed by the fact that my name appeared in several places at the Bienal, but there wasn’t really a relevant art installation in the space by me. I thought that was funny and beautiful.

Audioguide: more voices

There is an action in the 33rd Bienal project that involved a great deal of engagement with the Bienal’s maintenance workers (cleaners, assemblers, public monitors, etc.) that also provided very valuable information about the art system that was set up there. I’m interested in this worker archive. It’s important, relevant information – and unfortunately cultural institutions tend to neglect it. I’d love more people to know about the work of Iraildo Brito, the Bienal forklift operator – the only forklift driver in Brazil who can steer the vehicle around Oscar Niemeyer’s winding curves that characterize the Pavilion. This man is partly responsible for transporting the heaviest artworks to the upper floors of the building. It was he who transported Damien Hirst’s cow aquariums, for example. Why isn’t his knowledge treated as legitimate in the Bienal archive? Perhaps I’m more interested in what he thinks about the cows in formaldehyde than Hirst himself. In fact, I certainly am. At least in terms of materiality and space in relation to this work, Iraildo has a lot more to say than the British artist.

Technology as system

The technology that interests me comes from a very expanded notion of this concept [of technology]. To reinforce this breadth, I would choose to think of the idea of systems. Systems that are organized, often intricately, to be able to produce predetermined discourses. This construction is complex, as the discourse needs to be given a sufficient effect of neutrality to make it seem true, the almost spontaneous result of a certain instruction that has already been validated, pre-system. The less this effect of legitimacy is noticed, the better – it means that the system is up to speed. The organization of this, in whatever way and for whatever purpose, is part of the technology I’m interested in.

Machine Learning

That’s how I started researching the field of machine learning. I had previously been interested in the functioning of systems, such as the arts, and the organization of their elements to make the ‘magic’ happen. After my project at the 33rd Bienal de São Paulo, I was invited to a pilot artist residency in the laboratories of the then recently opened Artificial Intelligence Centre (C4AI) at the University of São Paulo’s Innovation Centre (C4AI, Inova USP). I created the Art and Artificial Intelligence Group (GAIA) there with the artist and professor Giselle Beiguelman, which allowed me to socialize with engineers, programmers, mathematicians, and so on. Little by little, I realized that the field of machine learning is also a large, complex system, organized in such a way that it has managed to absorb the words ‘intelligence’ and ‘artificial’ into its discourse, without the field actually affecting these elements. 

Image as truth

I think we’re in a process of intensification of something that has been happening historically: the image as map, instruction, a discourse of truth. The intensification of this is such that computer vision has only been able to advance through the use of millions of images that have historically been organized into databases. These are training images, or perhaps all images are. What could be different about this accumulation of training images? The awareness that images guide us. So, consciously, we can move on to a second chapter in this discussion. If images train, what do we want to train? This is the real power of the field of machine learning. How do we teach a machine? To what end?

Technophilia vs. technophobia

You have to strike a balance between technophilia and technophobia. Balance here shouldn’t be seen as sitting on the fence – because we know that big tech will completely control the actions that come from these warnings. If these warnings call for change, therein lies something important. I don’t think that artistic creation is primarily utilitarian – it’s even better if it’s not. But I do believe that in this case, art (in an expanded sense that has much more to do with exchanges, collective creative constructions than works of art and museums) can provide something practical: speculation. And, listening to the warnings, we will need this speculation to propose things, because we already know that the emptiness of this is the door to fascism. Just as art has shaped what machines are like today (their designs, modes, effects), it can offer more radical future ideas. What are we going to put in place of what we are criticizing? Therein lies a very promising opportunity for structural changes in the technological system of capitalism. At the moment, I’m obsessively thinking about it, as I’m conducting research that will culminate in a feature film about it, at the Collegium Helveticum – the Institute for Advanced Studies at the ETH, the University of Zurich and the Zurich University of the Arts. I’m working with several people on experiments that seek to speculate on new ways of organizing and aggregating information in images that train AI. I think arts research has a lot to offer (and radicalize) the field of computer vision.