Miha Turšič, Waag
Artificial intelligence is a mainstream term describing the scientific study of intelligence demonstrated by machines, in contrast to the natural intelligence of humans and other animals. Upon closer inspection we find that the sentiment surrounding popular AI discourses solidifies human exceptionalism – it is human intelligence that proves its primacy by demonstrating the capacity to create an aritficial version of itself.
This perspective however ignores numerous overlooked actors that contribute to the functioning of computational machines; in fact, we posit that the intelligence of machines is an environmental matter. The environment so substantially influences intelligence demonstrated by machines that we should understand it as an environmental condition in itself. Such a shift from artificiality toward environmentality allows us to address the blind spots of technological neutrality, which downplays technology's historical, social, and environmental embededness.
One of the main references of Waag's public research approach is Donna Harraway's notion that technology is not neutral[1]. In reference to her work, Chris Julien[2] argues that bias in AI is not a flaw in systems of measurement and automation, a 'remainder' to be exorcised, but a primary means of understanding, or rather accounting for our place and histories in the world, starting with our place in these artificially intelligent technological systems we talk so much about. Following bias, the dividing lines between domains such as technology and society, between nature and culture begin to diffract and collapse into each other, leaving us collectively stranded in the situation we once sought to discover, segment and control. Furthermore, the shiſt of perspective - from perfect windows allowing us to look 'onto' the world, to being 'in' the world together with these machines - directs our attention to the seams of AI implementation and calls for capacities to interpret the AI container.
While social, business, legal and ethical perspectives of AI dominate mainstream narratives, Waag applies an ontologically flat material view to focus on overlooked agencies embodied within the container of intelligence demonstrated by machines. In the Anatomy of an AI System,[3] Kate Crawford and Vladan Joler dissected Amazon's Alexa smart home device, mapping the controversial practices of rare metal extraction and invisible human labour, which are essential for the processing of digital tasks. The extraction of material resources was also addressed within the Supre:organism exhibition Waag organized with Vijfhuizen Kunsfort in 2019[4]. The Kongo Astronaut video work by Eléonore Hellio and Michel Ekeba concerned life on a planet made foreign by environmental and societal devastation following cobalt mining. Withing the same exhibition, the Space Offshore project by RYBN.org approached resource appropriation by showcasing extensive documentation on controversial outer space mining legislation in Luxembourg and the USA, which disregards international Space treaties. While technology providers usually remain disinterested to such concerns, artworks like these give the public an opportunity to reflect on the uncomfortable side of the status quo.
With environmental intelligence, the effort goes toward identifying the ways with which technology depends on and conists of environmental agencies. Planet Earth holds fascinating capacities expressed through chemical, biological, ecological and other processes. Science and technology that lead the effort of understanding the environment still recognize it merely as an object of observation or a resource, thus as something external. What artists observe is that through the processes of extraction and appropriation, the environment comes to live inside of us—humans and machines.
In Antti Tenetz's poetic and speculative work, Perihelion/Rage/secret_lover bacteria prosper solely on an analogue of lunar soil while deep dreaming, through neural networks, about what their life would look like on the Moon. We observe an organism that embodies the lunar environment and uses AI to daydream and speculate. Similarly, in PL'AI Špela Petrič developed a playground for the robot and plants to play. The artwork's neural network creates an abstract image of the plants as a mathematical matrix, just precise enough to identify the plants' movements so that the robot can reach closer to the tendrils that eventually hook up to the robot's interface. We observe neural networks imagining versions of an environment as relevant to the particular relationships in both cases. In this way, we again identify bias as a primary means of understanding both their and our place in the world.
Turning the gaze away from individual relations toward humanized environments, we not only observe the existence of bias but also its multiplicity. Aesthetics of exclusion with their project StreetSwipe[5]addresses the biased perception of an urban environment, specifically the aesthetics of gentrification. The artwork lets the audience determine if they think a photo of a bar storefront should be classified as 'gentrified'. Exploring digital environments, Tomo Kihara works with YouTube's recommendation algorithms in the TheirTube[6] project. The artwork demonstrates a trap of biased information bubbles where YouTube recommends users' popular content. In both cases, we observe a loop of influence, the way people as users form the physical and digital environment, while this biased environment locks its users into closed profiled containers.
As a way to shift the perspective away from human exceptionality in the context of intelligence demonstrated by machines, Waag advocates for the response-ability toward the material agencies expressed through emerging technologies like AI, for the diversity and inclusivity of interpretations in machine perception, and for transparent and participatory categorization of digital environments. With such an agenda, Waag challenges the environmental divide between technological and ecological environments. Poor environmental literacy usually passes from makers to their machines. With a focus on the expression of environmental agencies, the use of intelligence demonstrated by machines turns from very human interests to environmental care. In this way living systems larger than individual entities such as rivers, forests, oceans and clouds, can become recognised as active makers and users of intelligent machines.
[1] "A Cyborg Manifesto: Science, Technology, and Socialist-Feminism in the Late Twentieth Century," in Simians, Cyborgs and Women : The Reinvention of Nature (1991), pp.149-181.
[2] Chris Julien et al, Biased-by-default, position paper of AI Culture Lab, Waag
[3] Kate Crawford, Vladan Joler, Anatomy of an AI System: the Amazon Echo as an anatomical map of human labor, data and planetary resources, Novi Sad: Share foundation, 2018, ISBN 978-86-89487-13-8
[4] https://waag.org/en/event/supreorganism-exhibition