Jussi Parikka, Frédéric Neyrat, and Doron Darnov
Jussi Parikka is Professor at Aarhus University in Denmark where he leads the Digital Aesthetics Research Centre (DARC) and co-directs the Environmental Media and Aesthetics research program. He is also visiting professor at Winchester School of Art (University of Southampton) and at FAMU at the Academy of Performing Arts in Prague. His books include Insect Media (2010), Digital Contagions (2007/2016), A Geology of Media (2015), What is Media Archaeology? (2012), and Operational Images (2023). He is also the co-author of The Lab Book: Situated Practices in Media Studies (2022). His new book, with Abelardo Gil-Fournier, is coming out in June 2024: Living Surfaces – Images, Plants, and Environments of Media. He also works as curator, including as part of the curatorial team of transmediale 2023 and Helsinki Biennial 2023, as well as the co-curator of the Motores del Clima (Laboral, Gijon, 2023-2024)
1. Between What is Media Archaeology and A Geology of Media, you express a consistent interest in the intersection between media archaeology (the ways that extant media technologies historically emerge out of preceding ones) and the materiality of media technologies (especially in the context of metals, minerals, and mining). I wonder what opportunities AI might present for returning to this intersection. How do you understand the media archaeology and media materialities of AI? Can we fit AI under the familiar rubrics of thought that we might apply to investigations of more established forms of media technologies (for example: photography, film, computers, and smartphones), or does AI gesture to new forms of media materiality?
I would say that AI presents an interesting cross-section and parallel stream of impact, a peculiar recursive topic in its own right that does not fall into a normal category of “history of AI.” Let me try to explain.
On the one hand, it makes total sense to put current AI technologies and discourses into a historical lineage; this could concern for example automation but also the development of certain computational techniques as they emerge especially since the Cold War. This might then represent the media archaeological interest in understanding different timelines that are nested into earlier periods, a particular view to historical time that is not linear. A media archaeology of AI is a cross-section of automation, bureaucratization (as a particular history of formalized techinques of managing information), computation, etc. Also, it is not thus a history of mirroring human intelligence through an “artificial” model, but of different models of reasoning, optimization, problem solving and other operations that come to feature in a central way in different institutional practices.
On the other hand, the material history of AI is one of massive scales of supply chains and energy, harnessing resources for computational purposes. It is not just one “medium” but a large-scale infrastructure that can be discussed in different terms of scale. This runs parallel to the previous point, a sort of shadow history in its own right as it does not necessarily at first resemble anything “computational” and yet is part of the conditions of existence of advanced computation.
The question is how do these two lines intersect? I would claim this is part of critical research that is not merely reflective but also helps to glue the two together. Media studies that is interested in the materiality of computation (and other media) is as much needed as material thinking that is nuanced enough to see how any notion of media materiality is historically conditioned. Hence my point about recursion in this context.
2. In an infamous tweet, Ilya Sutskever recently declared that “it may be that today’s large neural networks are slightly conscious.” Might questions of “sentience,” “personhood,” or “consciousness” impact how we approach questions on the archaeology or materiality of AI? Or should we consider that such ideas might divert us from the materiality of IA?
I’ve always been interested in the movement that seems opposite to this claim: consciousness is perhaps slightly less interesting than some other forms of cognition, what Katherine Hayles coins the cognitive nonconscious. In other words, there are other forms of cognition that are not returnable to the usual models of sentience that seem often too narrow to understand “thinking” in the sense that more advanced cognitive studies let alone affect and critical cultural theory have proposed. This is why I pitch my Insect Media as a book about AI, animal intelligence. It concerns the legacy of brainless thinking, of such models of cognition and sentience that are about connectionism and coupling with the broader distributed milieu of agency. “This presents an entirely material way of understanding different forms of the embodiment of intelligence while avoiding the narrow anthropocentrism that creeps up in recent popular discourse: for example, when some hallucinate that large language models have become sentient.
So instead of personhood, I am interested in the material infrastructures of cognition and sensing, as well as in the questions of property or political economy that underpin actual media technologies and their discursive realms too.
3. You open a recent article, “‘A Natural History of Logistics’ and Other Studio Briefs: Problem Spaces for Planetary Design,” with the idea that “In recent years the planetary has become a recurring reference point across disciplines ranging from philosophy … to cinema and media studies, design, architecture, … as well as sociology.” What kind of conversations might AI begin to spark within this disciplinary turn towards “the planetary”? Might we (or, perhaps, should we) learn to think of AI as either a planetary technology or a technology of “planetary design”?
There are a couple of versions of this argument. The famous diagram by Vladan Joler and Kate Crawford mapped the material networks of AI systems, a sort of an anatomy that however did not turn inwards into the gut but to the various external networks through which AI is born as device, as service, as political ecology. And the other version is Benjamin Bratton’s take on planetary computation that includes related elements – Benjamin has been after all interested in the stack of computation from the earth to the cloud, via the city, interface, and user; this sort of planetarity is also patchy. Both concern fundamentally the question of scale of software as it intertwines with other socio-technological infratructures. It also concerns how the scale of design shifts from a sole focus on the interface to other kinds of spatio-temporal scales.
As a patchwork, planetarity is also in my view about the heterogeneous scales at which planetarity is not a given but produced across different techniques of mediation. AI is not merely born out of planetary materials but productive of the onto-epistemology of what we consider planetarity where we map resources, capacities, and regions in ways that becomes integrated into supply chains, for example. This is a material way of looking at geopolitics, as well as a geopolitical way of looking at media that underpin our sense of the dynamic volumetric space of planetary computation. My term “medianatures” was meant to frame this intertwining of material conditions and material epistemology of planetarity which is not necessarily of the scale of the planet imagined as a holistic entity.
So that text you mention looks at it from the perspective of studio practices and is interested also in method: what sort of methods of visualisation or conceptual thinking, architectural scales and other techniques make sense of the planetary that can thus be turned into a problem space (a hat tip to Celia Lury and her definition of the term) that involves the elemental media of planetarity: atmospheres, soil, etc.
4. We have a sort of open question concerning mapping and “alterity” (Spivak, quoted at the end of your article “‘A Natural History of Logistic”), open in the sense that your answer could be long or as short as a haiku: is there “something” that cannot be mapped? What might comprise this unmappable space of absolute, radical alterity (if such a thing exists!)?
I wonder if that what cannot be mapped we do not know yet? In other words, any radical alterity would then avoid this question and hence I am interested, in a somewhat Foucault-tone, in the “positivity” of knowledge and its material constitution; the map is the territory, to echo Bernhard Siegert. This does not deny the existence or validity of other knowledges!
So, to stay with the difficult question for a while, let’s stick with the word yet. Does the potential of radical alterity lie in the temporal dimension of potential virtual futures as much as it lies in potential virtual pasts? Hence, my earlier point contrasts with a more Deleuzian vibe. Of such knowledge formations – maps or media – that are not necessarily only fixing what we know into a grid but which allow for nomadic trajectories, of movement that itself produces the map at its wake. It is somewhere in this space where we also find many of the questions that decolonial and feminist scholars have successfully brought into view.
JOURNAL OF THE FIRST OUTERNATIONAL 