Meet the artists: Estampa
Estampa is a collective of programmers, filmmakers and researchers working in the fields of audiovisual media and digital environments. Together, their practice is based on a critical and archaeological approach to audiovisual technologies, on researching the tools and ideologies of artificial intelligence and on the resources of experimental animation.
How do you work as a collective?
Our practice as a collective doesn’t emerge from the collaboration of various artists, each one with their respective careers. We already were a team doing commercial work, mostly within the cultural sector. And, as a team, we have different profiles and everyone brings a particular set of skills to the table. But also having several interests in common, we wanted to explore them together, the same way we worked every day.
Thus, collective practice feels quite natural to us, even necessary. Decision-making is a shared responsibility, but if someone doesn’t have a clear opinion on something, or an urgent decision has to be made, we can comfortably delegate to the others. And if no one has a clear opinion, which happens… it probably isn’t that important of a decision.
In addition, working in the confluence of art and technology, the variety of skills and perspectives within the collective is especially valuable. The ability to test and develop some things by ourselves very much shapes our approach to making art and, although it might not necessarily translate into better results, it makes for a more immediate and flexible experimentation process.
How did you approach this new commission with Modern Art Oxford?
We are used to working with archives, or at least archive-like structures: collections, repositories, datasets… Even before we started working with AI tools, this has been one of our main interests, mostly related to the appropriation of existing materials. This kind of process is in part rooted in the desire to understand how and why these collections are constructed. So, there’s an implicit learning and there’s also a relation, through their work, with the people that made the original materials or that collected them. We like to say that it’s an archaeological approach.
This relation with others is usually, in our previous work, mostly related to their memory. In working with collections that have a clearer narrow focus, or that were collected in an automated way or by people with whom we don’t have any contact, the personal relation to others is not a key factor in the development of the work. But in the case of this commission, we are working with an archive that represents, in a very broad heterogeneous way, an institution with decades of history and from a foreign country to us. So, the relation with the institution, and the possibility to be there, has also been important. It’s possible that this fact has led us to reflect on the archive in a way that is maybe more tied to its present than its past.
There’s also a sense of motion when working with large collections, just by the fact that you cannot fully grasp them in a singular static form. You can arrange things in different ways, and you can establish all kinds of relations between different items. And when working with static images, you may see an instant of a particular movement that occurred, but by relating several of them you create a new sequence in which a new movement emerges.
To all of that, add the fact that you work with an archive in its physical form, and then the movement becomes literal. This is also something we haven’t done very much in our previous projects.
How did you use AI in your new commission?
The main use of AI within this work is for pose detection. That is, a procedure that identifies key points of each human body within an image in order to trace its simplified structure. This serves us as a way to highlight these bodies, but at the same time it is an abstraction that we can use to anonymise them.
Probably, the majority of images produced nowadays are run through some kind of artificial vision process, whether we are aware of it or not. Many of these images are not even seen by anyone ever –its sole purpose is to be run through these processes (see Paglen’s https://thenewinquiry.com/invisible-images-your-pictures-are-looking-at-you/). We can think, then, of a part of our contemporary visual culture, related to monitoring and surveillance, that is in fact not visible. Having AI as one of our main topics of reflection, we like to work with the visual representation of these processes. We want to place them again in the visible part of visual culture and we also find an interest in the aesthetics of their visualisation.
By using pose detection on the footage of our time in the archive, there is a tension between a process that operates in the digital realm and the physicality that gave sense to our activity there. As there is also a tension, we find, when using the aesthetics of monitoring to represent those actions, in a way that may serve as a bridge between the representation of cultural practice and the general notion of labour.
Additionally, using computer vision, we can explore the digitized contents of the archive in a way that allows us to find formal relations between images. On one side, we can select images based just on the similarity of the poses detected. On the other, we can arrange all images in a 2D map that attempts to put together those with closer formal features. This is a way of structuring the contents of the digital archive that may reveal interesting facts about them, or it may not. But, at least, it can be poetic. It is not a definitive arrangement by any means, but just one of the many possible. And it is always in confrontation with the actual, necessary, less flexible, physical arrangement of the archive.
Within your installation in Movements for Staying Alive, you explore the movements in our archive, from the practical tasks and the physical materials, to the projects documented. Could you tell us more about what interested you about considering the corporeal in the archive?
In our work with AI, we’ve worked a lot with datasets. These are the collections of data that are used to train AI systems: the examples of how a task that is being automated should be resolved. These are, obviously, digital. And each of the examples in a dataset is seen just as a singular piece of a more complex whole. But, at the same time, it is also isolated, in the sense that there’s no explicit relations between these singular examples. The fact is that a dataset is not meant to be explored, that’s not its purpose. This is in contrast with traditional archives, which, ideally, are meant to be consulted.
So, acknowledging the differences between what a dataset and an archive are meant to be, and sometimes subverting them, may be revealing of their contents and the collecting practice itself.
Exploring a dataset –that would be, treating it as an archive– is a very immediate way of getting some insight on it and, thus, on how AI is shaped. And using an archive as a dataset –training some AI system with it and observing the results– or even just analysing it with any digital tool, may help understand some of its idiosyncrasies.
But this latter approach means that the archive has to have a digital representation. For an archive that’s originally in a physical form, it has to forgo a digitisation process. The resulting digital representation may be convenient in portraying the contents of the archive, but it can’t be a complete representation of what the original archive really is. With the increasing digitisation of physical collections, this physical nature of the archive is becoming less present in the general conscience.
If we return to the dataset as a form of collection, we find, nonetheless, that this case is also true. Any data that’s a record of the non-digital world is a result of a digitisation process based on physical elements that exist or existed somewhere, most probably through the work of somebody. The way in which datasets are structured, presented and talked about most of the time doesn’t make this fact apparent. Having a particular task to automate in mind, superfluous data is stripped. The dataset, again, is not meant to be explored but automatically consumed. But the reality is that there’s a big labour structure involved in the creation of datasets. And the invisibility of this structure conceals some hideous practices regarding wage, worker rights and power relations between big tech and the global south.
Within this framework, we find that the relation between the body –the human body– and collected materials is a valuable tool to highlight both the physicality in which these collections are rooted and the notion of labour that surrounds them. Modern Art Oxford’s archive is ideal in bringing this relation forward, both in its form and in its content. Apart from its physical arrangement and the way in which archival and working space overlap, if we look at the images present in the archive, there’s a considerable amount of them that depict human activity. Some of them are related to performative practices, but most of them are related to interactions between the institution and its public. That is, a reflection of the work that the centre develops within a community.
Ahead of this commission, you spent time with us at the gallery exploring our archive. How did the experience of physically experiencing our archive translate into the digital and printed elements of your installation?
Right away, when we saw images of the space of the archive, we knew that this was of great interest to us. We couldn’t ignore the physical arrangement of it, and we wanted to be there. It’s not the same thing to peek at computer folders than to peek at boxes. And this action is directly represented in the installation. There’s footage of our own hands taking things from inside the boxes, and the record of our bodies moving through the room. But it’s not just us being there that interested us, we also wanted to get a hint of how the space is used to work. So, your bodies are in there too.
Apart from that, exploring the archive was also necessary to get a better sense of what it is. Seeing first hand all kinds of materials –both interesting and uninteresting– and how they are organised helped us understand the magnitude and scope of the archive (and the institution itself).
In the end, working within the archive reinforced our interest in its form and encouraged us to make this form more present in the final result, like having archive boxes in the installation or using post-it on the wall print. And there’s also a kind of parallel between the distribution of the installation and the distribution of the archive spaces: things arranged on the walls and a central working/generic space –the screen, in the case of the installation–, so you end up moving around the perimeter of the room. We have to say, though, that the characteristics of the exhibiting space and other practical things do have influence on this decision too.