Background 
Off
1 If you have enabled Personalized Ads, you can see your personal Google advertising profile here: adssettings.google.com. On Instagram, you can view the list of Ad Topics via your account settings, sorted by most recent interaction. Among the most curious topics in my top 20 list are “Men’s major golf championships,” “Racing,” and “Basketball Wives.” 

universal,
timeless,
celestial


Weiyi Chang


I visited Soft Turns in their studio in May. The studio was on the ground floor, a large, white-walled, open-concept space with a hard grey concrete slab floor. In the middle of the room was a model data centre they had built as a set for a new video they were in the process of shooting. The model was collapsible into modular units and reassembled on site; I had seen it earlier in the year as a pile of wood standing in the corner of their home. From the outside, it resembled a raised garden bed, with plywood planks stacked low to the ground; the inside consisted of long, narrow corridors flanked by dark-coloured computer blocks, dimly illuminated by unblinking strings of cold blue and white lights.

The artists draped a transparent plastic tarp over the model, transforming it into a greenhouse, albeit one that nurtured the growth and reproduction of data rather than vegetation. Blending the earthly, self-organizing logic of plants with the technological, algorithmic logic of computation, the gesture highlighted the similarities between the two spaces: both data centres and greenhouses are subject to precise and continuous climate control, both tend to be located outside of dense urban areas and both consume vast amounts of energy. Beyond their geographical and physical kinships, their similarities have also entered the cultural imaginary by way of metaphor. Consider the server farm, a predecessor to today’s massive globalized data centres, or the concept of data harvesting, which refers to automated processes that collect and synthesize data from various online sources and then use that data to advertise to you (among other, more nefarious activities1).

Elsewhere, the metaphorical relationship between agriculture and computation has become literal. In Boden, Sweden, there is a pilot project that combines data centres with greenhouses in a process called industrial symbiosis. The project will test the feasibility of using the massive amounts of heat and wasted energy produced by data centres to warm greenhouses in sub-arctic environments2. In Slovenia, some scientists have begun experimenting with plant DNA; by transcoding binary code into DNA sequences that can then be inserted into plants via bacterium, plants may be repurposed into stores of information, albeit in a read-only format suitable primarily for archival purposes.3

Would the Slovenian scientists have thought of using plant DNA as a potential data repository if we didn’t refer to DNA as a genetic code? The idea that DNA is a set of deterministic instructions, capable of being read, edited, encoded and decoded follows a long tradition of using mechanistic metaphors to describe more-than-human organisms. In a paper surveying the metaphors most commonly used in the field of synthetic biology, the authors identified three metaphor typologies: organisms as books that “can be read, edited and written;” organisms as engines or machines “that can produce ‘stuff’;” and organisms as computers that “can be programmed to do things.”4 The authors observe that each of these metaphors can be tied to major historical and technological revolutions, such as the invention of the printing press, the Industrial Revolution and the computer or information revolution.

2 Boden Business Park, “Large-scale greenhouse cultivation in smart industrial symbiosis,” ArcticToday, March 23, 2023. See also Cristina Ramos Cáceres et al, “Data-Center Farming: Exploring the Potential of Industrial Symbiosis in a Subarctic Region," Sustainability 14, no. 5: 2774.
Karin Ljubic Fister, “I plant memories in seeds,” New Scientist, January 16, 2016, 27.
Carmen McLeod and Brigitte Nerlich, “Synthetic biology, metaphors and responsibility,” Life Sciences, Society and Policy 13, no. 13 (August 2017).
Seb Franklin, “Cloud Control, or The Network as Medium,” Cultural Politics 8, no. 3 (2012): 458, DOI: 10.1215/17432197-1722154.
Ibid, 457-458.

The language we use to describe the natural and the technological world shapes our perception of reality; language informs the way we act upon the world, how we think about organisms and objects and the relationships we envision and create. This inevitably has ethical, social and political implications. The proliferation of mechanistic metaphors has facilitated nature’s exploitation, treating more-than-human organisms as automata condemned to their genetic and biological programming. Mechanistic and technological metaphors sanction a worldview that treats these entities as unconscious, unfeeling objects that can be controlled, re-engineered and exploited in the service of human interests. Concepts such as autonomy and self-determination, when applied to more-than-human entities, become moot.

The inverse is also true. When we use nature metaphors to describe mechanical and technological entities, we naturalize and neutralize the man-made systems and structures that increasingly delineate our lives. The concept of the Cloud is one metaphor that has come under scrutiny as of late. Clouds evoke images of weightlessness, universality, disembodiment and ethereality; clouds are seen as an endless natural resource that exists somewhere vaguely out there—a picture that is quite removed from the hardcoded material infrastructure that supports it. Writer and media theorist Seb Franklin argues that the amorphous, diaphanous concept of “the Cloud” functions as “representation of immateriality and smoothness that both effects and obscures the functions of a structured, striated grid that is the only representation of a world that is possible within the technical functionality of the digital computer.”5 The obfuscating power of metaphor has social and political repercussions: “Cloud computing represents the process, or at least the dream of a process, whereby the computer dissipates into an environment…It suggests the expansion of the cybernetic logic of informatic capture and definition to the status of periechon (that which surrounds) or atmosphere, the theoretical emergence of both the world-as-computer and the computer-as-world.”6

Courtesy of Soft Turns

Franklin’s fears about the “expansion of the cybernetic logic of informatic capture” have intensified in the wake of popular artificial intelligence tools like OpenAI’s GPT-4 and Meta’s LLaMA. GPT-4 and LLaMA are large language models (LLMs) that have been trained to recognize, translate, predict or generate text, code and other content using massive data sets. The initial excitement prompted by the launch of these LLMs came under scrutiny recently when it was revealed that the models had been trained on copyrighted literary works.7 Similar issues have arisen around AI text-to-image generators trained using copyrighted works of art. While LLMs and AI might not be capable of mimicking genuine human creativity just yet, that hasn’t stopped corporations from abusing AI’s potential to further devalue the work of writers and artists.8 

Not only are these programs trained on vast quantities of data, they also consume massive amounts of energy—given the state of clean energy adoption globally, the energy is mostly sourced from coal, oil, gas and other non-renewables—to keep their servers operational. A single data centre can consume as much electricity as 50,000 homes; 40% of the energy used is spent on cooling alone. When we add the electricity consumed by data centres to the energy used by networked devices (laptops, smartphones, smartwatches, tablets and all manner of devices that constitute the Internet of Things), the total electricity consumed by the Cloud amounts to 2% of global carbon emissions.9 Regrettably there are few if any regulations governing data centres’ energy and water usage.

Furthermore, despite their environmentally-friendly rhetoric, few companies building and operating such facilities are invested in developing energy- and resource-efficient data centres. In conversation with a former classmate and architect who designs data centres, he revealed that, despite his employer’s highly publicized commitment to achieving global carbon neutrality within the decade, energy efficiency concerns were nearly non-existent on the list of priorities, trumped by more immediate concerns including the availability and number of utility lines, the quality of the water table and water access, and whether or not the site’s governing body was willing to offer discounted energy rates.

7 Alex Reisner, “Revealed: The authors whose pirated books are powering generative AI,” The Atlantic, August 19, 2023 To see which books were included in the dataset, see Alex Reisner, “These 183,000 books are fueling the biggest fight in publishing and tech,” The Atlantic, September 25, 2023.
7 Consider, for example, the recent SAG-AFTRA and Writer’s Union strikes. Natalie Jarvey, “SAG-AFTRA Strike: Studios’ AI Proposal Sounds Like Black Mirror, Right?Vanity Fair, July 13, 2023. See also Derrick Bryson Taylor, “Tom Hanks Warns of Dental Ad Using A.I. Version of Him,” The New York Times, Oct. 2, 2023. For a comedically prescient take, see Tina Fey’s 2007 30 Rock season 2, episode 1, “SeinfeldVision,” in which NBC executive Jack Donaghy (portrayed by Alec Baldwin) proposes inserting NBC-owned footage of old Seinfeld episodes into current programming.
9 Steven Gonzalez Montserrat, “The Cloud is Material: On the Environmental Impacts of Computation and Data Storage,” MIT Case Studies in Social and Ethical Responsibilities of Computing no. Winter 2022 (January).

But if metaphor can veil ethically dubious behaviour and mask social and political repercussions, it can also be used to posit alternative futures. Soft Turns’ experimental data centre-greenhouse hybrid gives form to the possibilities and promises latent in the amalgamation of technology and nature. In this respect, Soft Turns’ approach to industrial photography differs from the documentarian impulse of predecessors like Bernd and Hilla Becher, whose photographic assemblages recorded the typologies of obsolete industrial architecture, or contemporaries like Edward Burtynsky, who uses photography to capture the devastating impact of industry on landscapes.

Courtesy of Soft Turns

While the Bechers and Burstynsky look back at the industrial landscape, reflecting on the scars and structures that inscribe capitalism’s interventions in the land, Soft Turns look ahead at what is yet to come. In line with their broader methodological approach, Soft Turns set forth a hypothetical proposition: What if we think of the data centre, not as a mere industrial facility, but as an ecosystem? What kind of image-world is possible within the confines of the data centre? What kind of environment might emerge if computation was left to its own devices?

For Soft Turns, metaphor opens up a space capable of holding and generating speculative futures. The greenhouse is not a literal greenhouse to be appended to a data centre, but rather evokes the warming planet we now inhabit, a planet whose existing climate regulating functions are on the brink of collapse. The data centre and the Cloud it supports becomes a monument to humanity’s hubris, a structure that claims to unlock an infinite space unbound by material restrictions, an ideal world in which everything can be captured, controlled, analyzed and aggregated without limit. As we enter a new technological era dominated by AI and machine learning, what new metaphors will emerge that will reshape the ways in which we conceive of the natural world? And, conversely, how will a warming planet, an inconceivable loss of biodiversity, a natural world in chaos, force us to rethink the scope and direction of technological development?


As we sat on the concrete studio floor, I was reminded of Alan Weisman’s 2007 book, The World Without Us. Weisman’s thought experiment asked what would happen to the man-made artefacts we have wrought upon the planet if humanity simply disappeared. The book is segmented into chapters that address a different type of human artefact—houses, cities, art, plastic, farms, nuclear facilities—and how they might fare without the aid of human labour perpetually warding off entropic decay. Weisman paints a vivid picture of nature’s reclamation, the futility of our monuments to ourselves, as well as the toxic legacies we leave behind.

As far as I can recall, Weisman does not write about data centres as such. Some of the materials might persist—chromium alloys used to make stainless steel, ceramic tile floors and glass windows, plastic polymers. But in a world that is warming faster than at any other point in the geological record, as the energy grid strains under unyielding demands and as the backup generators run out of fuel, as groundwater is depleted and poisoned and as the cooling systems collapse and infiltrate the servers, the Cloud—universal, timeless, celestial—will share the same fate that awaits the rest of us here on the ground. 










Bibliography

Boden Business Park. “Large-scale greenhouse cultivation in smart industrial symbiosis.ArcticToday, March 23, 2023. .

Bridle, James. “Under the Cloud.BBC4, October 13, 2020.

Cáceres, Cristina Ramos, Suzanna Törnroth, Mattias Vesterlund, Andreas Johansson, and Marcus Sandberg. 2022. "Data-Center Farming: Exploring the Potential of Industrial Symbiosis in a Subarctic Region." Sustainability 14, no. 5: 2774.

Fister, Karin Ljubic. “I plant memories in seeds.” New Scientist. January 16, 2016.

Franklin, Seb. “Cloud Control, or The Network as Medium.” Cultural Politics 8, no. 3 (2012): 443-484. DOI: 10.1215/17432197-1722154.

Glanz, James. “Power, Pollution and the Internet.The New York Times, September 22, 2012.

Hogan, Mel. “Environmental media” in the cloud: The making of critical data art.New Media & Society, vol. 25, no. 2 (Feb. 22, 2023).

Jarvey, Natalie. “SAG-AFTRA Strike: Studios’ AI Proposal Sounds Like Black Mirror, Right?Vanity Fair. July 13, 2023.

McLeod, C., B. Nerlich. “Synthetic biology, metaphors and responsibility.” Life Sci Soc Policy 13, 13 (2017).

Monserrate, Steven Gonzalez. “The Cloud Is Material: On the Environmental Impacts of Computation and Data Storage.MIT Case Studies in Social and Ethical Responsibilities of Computing no. Winter 2022 (January).

Pratt, Mary. “Cloud computing's real-world environmental impact.TechTarget, June 7, 2023.

Reisner, Alex. “Revealed: The authors whose pirated books are powering generative AI.The Atlantic, August 19, 2023.

Reisner, Alex. “These 183,000 books are fueling the biggest fight in publishing and tech.The Atlantic, September 25, 2023.

Taylor, Derrick Bryson. “Tom Hanks Warns of Dental Ad Using A.I. Version of Him.The New York Times, Oct. 2, 2023.

Weisman, Alan. The World Without Us. New York, US: HarperCollins, 2007.


  1. If you have enabled Personalized Ads, you can see your personal Google advertising profile here: adssettings.google.com. On Instagram, you can view the list of Ad Topics via your account settings, sorted by most recent interaction. Among the most curious topics in my top 20 list are “Men’s major golf championships,” “Racing,” and “Basketball Wives.”
  2. Boden Business Park, “Large-scale greenhouse cultivation in smart industrial symbiosis,” ArcticToday, March 23, 2023. See also Cristina Ramos Cáceres et al, “Data-Center Farming: Exploring the Potential of Industrial Symbiosis in a Subarctic Region," Sustainability 14, no. 5: 2774.
  3. Karin Ljubic Fister, “I plant memories in seeds,” New Scientist, January 16, 2016, 27.
  4. Carmen McLeod and Brigitte Nerlich, “Synthetic biology, metaphors and responsibility,” Life Sciences, Society and Policy 13, no. 13 (August 2017).
  5. Seb Franklin, “Cloud Control, or The Network as Medium,” Cultural Politics 8, no. 3 (2012): 458, DOI: 10.1215/17432197-1722154.
  6. Ibid, 457-458.
  7. Alex Reisner, “Revealed: The authors whose pirated books are powering generative AI,” The Atlantic, August 19, 2023 To see which books were included in the dataset, see Alex Reisner, “These 183,000 books are fueling the biggest fight in publishing and tech,” The Atlantic, September 25, 2023.
  8. Consider, for example, the recent SAG-AFTRA and Writer’s Union strikes. Natalie Jarvey, “SAG-AFTRA Strike: Studios’ AI Proposal Sounds Like Black Mirror, Right?” Vanity Fair, July 13, 2023. See also Derrick Bryson Taylor, “Tom Hanks Warns of Dental Ad Using A.I. Version of Him,” The New York Times, Oct. 2, 2023. For a comedically prescient take, see Tina Fey’s 2007 30 Rock season 2, episode 1, “SeinfeldVision,” in which NBC executive Jack Donaghy (portrayed by Alec Baldwin) proposes inserting NBC-owned footage of old Seinfeld episodes into current programming.
  9. Steven Gonzalez Montserrat, “The Cloud is Material: On the Environmental Impacts of Computation and Data Storage,” MIT Case Studies in Social and Ethical Responsibilities of Computing no. Winter 2022 (January).

Weiyi Chang (she/her) is an independent writer and curator. Weiyi has curated exhibitions and programs in Canada, the United States, and Germany. Her art criticism and essays have been published in Canadian Art, C Magazine, and Luma Quarterly and she has contributed to numerous exhibition catalogues published by the Whitney Museum of American Art, Documenta 14, Morris and Helen Belkin Art Gallery, and more. Weiyi's research lies at the intersections between ecology, environmental ethics, climate change, capitalism, and time.

Weiyi was a 2019-20 Helena Rubinstein Curatorial Fellow at the Whitney Museum of American Art's Independent Study Program. She holds a MA in Art History (Critical and Curatorial Studies) from the University of British Columbia and a BA (Honours) Major in Art History and Major in Philosophy from Western University.

Related Content