On Separating Data from People: Reflections on Indigenous Digital Exhaust, Transparency, and Constitutive Relationality

ABSTRACT

This essay invites readers to think through the carelessness and care-lessness of data extraction regimes under surveillance capitalism with regard to notions of care produced with and against these systems. It does so by first narrating Joanna Radin’s analysis of how the Akimel O’Odhum came to be separated from their data as “exhaust,” despite the indigenous data sovereignty movement. It then shifts registers to spotlight the kinds of care evident under datafied logics of transparency in organic cotton supply chains in the Global South, contrasting it with the holistic and constitutively relational care articulated by indigenous communities who produce cotton. It concludes by arguing for the ethical and political urgency of enacting practices that center the relation and connectedness between people, place and data.

KEYWORDS:

• Data
• indigeneity
• transparency
• digital exhaust
• constitutive relationality

Care, as has been well documented by indigenous scholars, feminist critics, and activists (Hobart & Kneese, 2020; Pezzulo, 2023), is at the center of the moral imperative to act in the face of the multiple interlinked inequities, violences and degradations that characterize contemporary capitalism. For those of us invested in environmental justice, resource extraction is a central factor in such degradations, and increasingly, such systemic extraction involves and, in fact is centered around data. Digital data themselves seem particularly innocuous because of their apparent virtuality, but their materiality is now underscored by the fact that digital infrastructures now account for nearly 4% of global greenhouse gas emissions (Dourish, 2017).¹ Surveillance capitalism itself is composed of such digital material, upon which elaborate infrastructures of data extraction, dispossession, rendition and ultimately, prediction, are built (Zuboff, 2019).

And so, in this essay I attend to the central question of how we might think through care as it is squeezed out of, reinserted into, and ultimately excluded from global systems of data extraction. I do this by taking the reader through three interconnected narratives to illuminate the carelessness and care-lessness of data extraction regimes under surveillance capitalism with regard to notions of care produced with and against these systems. The first narrative builds on Joanna Radin’s account of how the Akimel O’Odhum came to be radically separated from their data, to illuminate the central role of “digital exhaust” in shaping the carelessness and care-lessness of machine learning enterprises. The second narrative draws from my fieldwork on organic cotton supply chains that stretch from India to Sweden to shed light on the nature of care produced by datafied transparency logics that continue to separate people from data. The third narrative centers the voice of Machamma, a tribal cotton grower in Kabini, South India, who articulates a deep notion of care grounded in a constitutive relationality between land, plants and people. I conclude by arguing for the need to reckon critically with the separation of people and data as we develop conceptions of care to guide justice praxis.

On the care-lessness of digital exhaust

Joanna Radin, a historian of science, recounts in wonderful detail a conversation she had with a mathematician who studied machine learning. Their story was about working with a team of mathematicians that fed data into a predictive algorithm to help the underground New York City electric grid become more fire resistant (Radin, 2017a). None of this data was collected by the mathematicians themselves; rather, they used preexisting data sets, a common practice in computational machine learning.

One of these data sets, it transpired, was colloquially called “Pima,” which Radin established had been collected over a period of decades from the Akimel O’Odhum, a Native American community who were settled in the Gila River Indian Community Reservation outside Phoenix, Arizona. The formal name for the Pima dataset was PIDD – the Pima Indian Diabetes Dataset. This dataset comprised information about diabetes prevalence amongst Akimel O’odhum including age, gender, pregnancy, body mass index, insulin levels, blood pressure, and diabetes pedigree. In short, the data were digital representations of the very bodies of Akimel O’odhum.

Data collection about Akimel O’odhum, Radin says, was really begun by the anthropologist Frank Russell in 1901, 40 years after the establishment of the Pima River Indian Community Reservation. Russell himself died of tuberculosis two years later, aged 35, but not before he apparently identified just one case of diabetes mellitus in the community. More to the point, his was the first systematic effort, funded by the Bureau of American Ethnology, to collect data about Akimel O’odhum that highlighted their relative deprivation, an early instance of a deficit discourse model of data collection about indigenous communities all over the world (Fogarty et al., 2018).

Fifty years later, after decades of government “assistance,” diabetes had been established as a major scourge in the community, and in 1963, what scientists called “cooperative” research between Akimel O’odhum and the National Institute of Diabetes and Digestive and Kidney Diseases began. This arrangement launched the longitudinal studies that became PIDD in the 1990s. As Radin notes, there was no discernable improvement in either obesity or diabetes rates in the community during the 40 year study, despite a prevention program launched in the 1980s. Further, it was only in the 1990s that community members began receiving compensation for their participation in the study, including health care, in the form of 50 dollars. Activism around Native American health and land issues and their concomitant visibility increased considerably during this time (Estes, 2019).

Geneticists have over the years made several arguments for the value of the PIDD dataset, ranging from the now vilified “thrifty genotype” hypothesis advocated in the 1960s by James Neel,² who argued outlandishly that because indigenous genes were more used to deprivation and adversity they would not cope with modern diets, to the more accepted idea that the genetic homogeneity of this pool made identifying the causes of diabetes mellitus that much easier. These latter arguments in fact facilitated the entry of PIDD into the University of California at Irvine Machine Learning Repository. Started in 1987 by David Aha and his graduate students, the Irvine repository was a result of a call by several computer scientists in major universities across the United States. Various datasets were donated rapidly over the next several years to the repository with various conditions, mostly tied to attribution to collectors and funders rather than to patients and the sources of data themselves. The PIDD dataset was “donated” to the repository in 1990, and Radin says it was one of the earliest datasets. Its donor was listed as Vincent Sigillito, one of Aha’s mentors at Johns Hopkins University.

Radin says that PIDD had been accessed close to 260,000 times by 2017. My own examination found that the dataset contains no information as to whether or not Akimel O’odhum consented to making their data available or not, and it is clearly the case that they never did benefit from its creation. Radin recounts how in 2009 the Gila River Community enacted an ordinance that says that “Medical and Health Care Research has been conducted in ways that do not respect the human dignity of human subjects and that do not recognize the legitimate interest of the Community in the integrity and preservation of its culture” (p. 58). Akimel O’odham have closed their doors and are no longer talking to medical researchers. And since about 2021, PIDD has been removed from the Irvine repository.

Despite this obvious enactment of a politics of refusal (Simpson, 2014) by Akimel O’odham and the removal of the database from the repository, PIDD is actually everywhere else in 2023 – on Github, on Network respository.com, and on Kaggle, where someone named “Kumar” is listed as the owner. In 2022 and 2023 alone, it appears to have been cited many hundreds of times. My review of 23 of these samples revealed that while all these studies cited Irvine, none of them even mentioned Sigitello nor provided a single sentence contextualizing the dataset containing information about the bodies and health of over 3000 mostly dead members of Akimel O’odham.

This circulation of data as an autonomous property, alienated from the Indigenous community from which it was extracted seems like carelessness writ large. It is clear that these researchers do not feel the need to account for or even consider the problematic origins of the data that they feed into their machines. On the contrary, the data is considered to be both ubiquitous, i.e. freely available because of prior research and what Radin calls “naturally occurring—a neutral product of the contingent circumstances of its acquisition” (p. 58). Further, precisely because the data is decontextualized – recall that it was used in machine learning about a fire drill – it is easy to forget that it is built on top of pain, disease and a brutal history of settler colonialism.

But is the casual, detached and abstracted circulation of Akimel O’odham’s data fully described by the word “carelessness?” Disguising the origins of data also disguises the extractive and predatory, in fact, parasitic nature of the machine learning enterprise itself, which at least at the surface, implies that if we do not know, it does not matter, and that all we are doing is collecting digital exhaust. The latter is itself a problematic term. I can trace the beginnings of its use to describe a particular kind of data back to the year 2009, when a few researchers began to use it to describe meta-data that were developed and constructed from digital transactions ranging from business processes to GPS tags and calendar entries.³ It has since become an important term in scholarly accounts of the machine learning apparatus (Leonardi, 2021). The exhaust metaphor, grounded in an environmental pollution imaginary, implies, misleadingly, that such data are unnecessary byproducts of more “real” digital transactions rather than the intentional and systematic data capture by a digital enclosure: as Mark Andrejevic wrote tellingly, a generation ago, of such enclosures, every act of communication is designed to generate data about itself (Andrejevic, 2007). Technology critics such as Ruha Benjamin remind us, machine learning as a practice is designed to be parasitic: it is not just careless or agnostic about data, it needs data to be acontextual, irrelevant, abstract, anonymous and intrinsically dehumanized, so that it can monetize human experience (Benjamin, 2019). Digital exhaust is ultimately not so much careless as it is care-less, drained of and unable to admit care into its framework by virtue of its very constitution.⁴ Read in this way, big data might be imagined as the ultimate perversion of the anonymity clause upon which ethical research and practice ought to be conducted.

It is for these reasons that work on indigenous data sovereignty has gained traction considerably in the last ten years, and now represents an important frontier in protecting the exploitation of traditional knowledge in the name of the commons. A collective of indigenous researchers, based mostly in Aotearoa New Zealand, Australia, Canada and the United States, have been arguing since at least 2014 that indigenous communities need to assert and exercise sovereignty on their data for reasons that Radin’s archival work makes clear, naming specific sets of principles for data management and policy that enshrine indigenous rights. Key activist-scholars in this movement include Tahu Kukutai, John Taylor, Maggie Walter, Stephanie Caroll Rainie, Rebecca Tsotsie, Michele Suina, Desi Rodriguez Longbear, Maui Hudson, and many others.⁵ They have also created activist-scholar networks based on the idea that data are people too. Three national-level indigenous data sovereignty networks are particular active: the US Indigenous Data Sovereignty Network, Te Mana Raraunga Maori Data Sovereignty Network in Aotearoa New Zealand, and the Maiam nayri Wingara Aboriginal and Torres Strait Islander Data Sovereignty Group in Australia. The Global Indigenous Data Alliance in 2020 published a set of principles for indigenous data management, tellingly titled with the acronym “CARE” to emphasize collective benefit, authority, responsibility and ethics in managing indigenous data (Global Indigenous Data Alliance [GIDA], 2020).⁶

Such scholarship and activism are an indigenous foil to the work of the Care Collective,⁷ and their studies of indigenous resistance can be seen as attempts to decolonize the seeming openness of data as acts of resistant care. Maggie Walter and Michele Suina, for example, have documented the work of the Albuquerque Area Southwest Tribal Epidemiology Center (AASTEC) from 2006, which advocates for tribal determination of what data instruments can capture, where data goes; what it is used for; and the need for the center to work as an intermediary between the federal government and tribes, to facilitate the asking of the return of data from scientists (Walter & Suina, 2019). Others have made visible the 2002 Navajo Nation’s “Banishment Order” to prevent collecting and contributing samples and genomic data. As communication researchers interested in environmental and human care, we should be deeply invested in learning lessons that this scholar-activist movement is teaching us (Garrison, 2013). Certainly, it helps us reconfigure how we understand human risk in research and the damage we do when we separate people and data. We all need to fear that big data, because it radically transcends the circumstances of its production further radicalizes the separation of both people from their data as well as data from their people.

Care in and around organic cotton supply chains

In addition to my recent research on indigenous databases, over the last four years I have had the opportunity to lead fieldwork into communicative dynamics that shape and surround organic cotton supply chains that stretch from farms in India to retail outlets in Sweden.⁸ My work here has been driven by a project funded by the Swedish Foundation for Strategic Environmental Research, and along with a team of researchers based in five countries, I have been studying how transparency is produced in these chains. Transparency in supply chains has often been described as happening when experience turns into data, i.e. when it is inscribed, disclosed, and made visible (Harness et al., 2022). Transparency, I posit in this context, is constructed as the ethical foundation upon which the separation of people from their data becomes possible. In organic cotton supply chains, transparency is managed through one of two processes: (1) testing, to ensure that products and commodities meet standards and are in fact organic; and (2) traceability, where data about products is available such that companies and ultimately consumers are assured that their products are in fact coming from organic fields and have been made according to the ethical standards prescribed by certifying bodies and sustainability experts in the Global North.

Our investigations have helped uncover multiple meanings of care in such systems; one inextricably caught up in transparency logics itself, and the other less visible, more holistic, and deeply, even constitutively relational. Over the last few years, we have engaged with retail outlets, clothing companies, garment factories, ginning mills, organic cotton farms, and handloom centers, interviewing managers, designers, garment workers, weavers and farmers, trying to make contact with as many links in the chain as we can. For actors in the chain, making it transparent was, we found, was seen as an act of environmental care that contributed towards sustainability. The companies we worked with in Sweden practice what some scholars have called a kind of radical transparency (Heemsbergen, 2021) – throwing open their doors to us with the belief that if they make their processes, values and work known, that would result in better, more sustainable and more ethical practice. Radical transparency as an organizing practice became more commonplace in the garment industry after the tragic Rana Plaza accident in Dhaka in 2013, where a building housing five garment factories that exported clothes to the Global North collapsed and over 1100 workers working on subsistence wages perished (Ganesh & Rudrappa, in press).

The two companies we engaged most locate their sustainability efforts tellingly, in their production teams rather than their marketing teams. Their sustainability managers visit India often, going to the very same locales that we visited; and in turn, they strive for deep engagement. One of these managers even emphasized to us several times that they are conscious of not replicating colonial modes of production, as they are interested in ensuring that livelihoods of workers in the supply chain are as sustainable as their products. These managers appear sincere, committed to a discourse of ethical capitalism (Barry, 2004),⁹ and provide a valuable corrective to the excesses of unsustainable fast fashion that is now driven globally by the US market (Niinimaki et al., 2020). They have worked hard to, as they put it, know their supply chain- to be familiar with where the material they use comes from, with conditions of manufacture and production, and with the farmers themselves. From my observations, the notion of care they articulate is deep, resonant and heartfelt. Two vignettes from our fieldwork, however, interrupt the smooth incorporation of care in such narratives of sustainability and transparency emerging from the Global North side of the supply chain.

Care on the shop floor: a logic of transparency

In 2022, I had the good fortune of being able to take nine students from our US public university to Sweden and India to learn something about the work our international research team had been doing. As soon as we got to Bangalore, we got into a bus and drove a few hours south to the factory where garments were made to send to Sweden. We toured the impressive facilities and were taken around the premises by the director himself, a charismatic and committed individual who had started the factory a few years earlier, working exclusively with organic cotton, drawing from networks across the country that he had helped build. He saw it as his mission to employ local workers rather than rely on migrant labor, reaching out particularly to women who were single parents, especially from Dalit communities, to ensure that they were able to stay in their villages and in their extended family networks rather than have to migrate to Bangalore or Chennai to perform domestic labor, more exploitative garment work, or sex work.

As we toured the factory, we were struck by what was clearly high-quality industrial infrastructure and work processes that were designed carefully through time and motion studies. These assembly lines to produce different kinds of shirts for a range of high-end clients around the world were constructed painstakingly, calibrating down to the last millimeter how cloth was inspected, measured, cut, stitched, finished, ironed, and packaged. Each production line consisted of between 13–15 women and there were about seven lines operational at any given time. In line with their transparency requirements, the entire process was carefully documented, and the data shared and verified with each individual client in Europe to make sure that orders were happening to specification, that quality standards were being met, and that work was being done in a timely way. At the same time, European clients were provided carefully audited data about wages workers were being given, how wage levels were calculated, how many hours they worked, the demographics of the labor force, and even promotion practices and absenteeism rates.

Our students engaged with workers on the assembly line enthusiastically and were moved by workers’ willingness to engage with them honestly, happily and sincerely. One moment was particularly poignant. A worker revealed that while she was happy at work, her life’s ambition was to learn how to make an entire shirt from start to finish. She wondered who wore them, what they did while they were wearing them, and what the life of the shirt ultimately was. This short interchange, and this yearning to know stood in sharp contrast with the disciplinary efficiency of the production line and the overall docility of the workers bodies. It made visible, just for a moment, the deskilling and ultimately alienation involved in the construction of an assembly line. It momentarily threw into relief the mundane or maybe even dreary task, that took a considerable physical toll, of having to stitch a collar onto a shirt every eight or nine minutes, and repeat it day after day until the production quota was reached, the product of labor disappeared, line disassembled, and a new one formed. Most poignantly, it demystified the commodity itself by laying bare the subjectivity of the many workers who together made the shirt. These subjectified yearnings and desires do not detach themselves from people and turn into data that the factory produces and that are sent to Europe. They remain on the shop floor, quite possibly discarded at the end of the day, when the (predominantly) female workers go home generally to care for their children and families.

Care on the shopfloor here, remains caught up in a logic of transparency that safeguards basic building blocks of capitalism: the routinization of worklife, the deskilling of labor, and the preservation of global hierarchies of commodities and capital. Logics of transparency ensure that acts of care themselves, whether it be inclusive hiring, gender and caste-sensitive employment practices, or support during natural and health disasters are seen in terms of benevolence rather than as fundamental components of the system of production.¹⁰ Here, managerial care is thus only one element of a broader set of workplace transactions. It stands in contrast with how foundationally garment workers position care in their own lives and why they come to work in the first place: the need to care for their families and children.

In the brief conversations we were able to have with garment workers about their families, I sensed a duality. On one hand workers expressed gratitude for the work they did and the relatively benign conditions of their employment, the factory was clearly a valued employer in the community, and they respected and liked management and the charismatic founder. In fact we observed lines of women outside the factory gates in the morning waiting for announcements about possible vacancies, and we know that the turnover rate in the factory is less than 5% a year, which is remarkable for the industry. On the other hand, most women expressed a clear preference for a different future for their children, one not attached to the factory, but that often involved education, a move to a city, or a government job, rather than working as they did in a factory in a small provincial town. Caring futures did not, for them, involve garment work (Rudrappa, 2015).¹¹

Care as constitutive relationality

Care as a fundamental element of being in the world was evident in fieldwork that I did with my colleagues Kashika, Srikantha and Maada in and around Bandipura and Kabini in the Nilgiri Biosphere. We had been visiting tribal hamlets around the perimeters of a national park, to engage with formerly nomadic indigenous communities who had been displaced and resettled into these hamlets when the park was created and they were no longer able to access their ancestral lands. Some community members had been allotted small parcels of land as part of that displacement and had been persuaded (or elected) to grow cotton. In one particular round of fieldwork in late 2021, we went to about eleven villages around Kabini. Here, we registered and witnessed a similar sense of resignation that we had encountered in Bandipura: farmers were sometimes stoic, sometimes passive and sometimes sad about the fact that they had to farm at all in the face of their displacement from the forest. There were worse fates, some of them said, even as they showed us around their small acre-large plots of land. The original displacement from the forest, we felt, had produced a range of related affective stances- some were phlegmatic, others even accepting – but for all of them, the severance from the forest still loomed large even if they had been born in the resettlement colonies. We were also struck by the fact that they only appeared to be partially integrated into the logic of a cash economy, as evidenced by their lack of means-ends calculation when it came to figuring out the price of seeds, labor, water or organic fertilizers. In fact, only one family we visited was actively thinking of producing cotton in terms of investment and maximizing returns.

In one village in Kabini, Kebbebura, we met and conversed with Machamma. Her family was of the Jenu Kuruba tribe. Along with 200 other families they, had been displaced to Kebbebura some 60 years previously, and she herself had been born in the forest before leaving it as an infant. Machamma was very conversant with popular techniques of growing organic cotton, and also other crops like Ragi, Jowar and horse gram. She had been growing cotton for about ten years, and spoke knowledgably about making organic fertilizers for it, which, as she put it, would “cool the land” given the increasing heat they had been facing over the last decade (which is one reason she switched to cotton cultivation in the first place). She mentioned that others had tried to get her to grow conventional cotton because it was less expensive, but she resisted because she saw value in what she did. In fact, one of her messages to the outside world was “what we are doing here, you can also do it, you can learn. Everyone will live well and happily if we have good health.” She went on to relate this view to her connection to the forest.

What we have learned is that we must not abandon the forest. The forest needs honey to be harvested, yam needs to be eaten, so it can come up again, and water needs to flow, so more water can fill [the watering holes]. If there are a hundred mushrooms or bamboo shoots, we should use them and not waste them; the forest and us will both be well if we do that, and our health will be good. That is our method.

In this way, she connected the production of organic cotton not with its market value, but with a discourse of moderation and a deeper symbiotic connection between land, plants, animals and people, as well as the need to try to continue traditional methods of gathering, growing and regenerating to preserve these fundamental relationships. Thus, for her care for the crops was a constitutive relation that could not be separated from care for land, plants and other people (Kimmerer, 2013).¹² Such holism as a form of care was utterly foundational to her way of life, and in this light, her subsequent lamentation about not being able to freely return to the forest anymore was devastating: “we want to live in the forest: that is our only wish.” Her practice of care stands in contrast to the epiphenomenal place that transparency-oriented and managerial versions of care occupy vis-à-vis the organic cotton supply chain, where assembly line production continued to proceduralize, informationalize and deskill workers, even in a context grounded in sustainability, where workers are relatively well off.

What I am calling “constitutive relationality” in this essay resonates with and is inspired by Dwayne Donald’s (2009) discussion of ethical relationality produced as a result of deeply hybrid indigenous practice, or Indigenous Métissage (Donald, 2009). Donald describes ethical relationality thus:

Ethical relationality is an ecological understanding of human relationality that does not deny difference, but rather seeks to more deeply understand how our different histories and experiences position us in relation to each other. This form of relationality is ethical because it does not overlook or invisibilize the particular historical, cultural and social contexts from which a particular person understands and experiences living in the world. It puts these considerations at the forefront of engagements across frontiers of difference (p6, emphasis mine).

Concluding reflections

What does it mean to, at least provisionally put these narratives from different hemispheres together, and what do they say about care? The first narrative illustrates the limits of ordinary carelessness as a metaphor for thinking about data extraction and its severance of people from place; instead, it helps us draw attention to the intentional and designed character of contemporary modes of data extraction that can only work by separating data from context, thereby ensuring that care does not even enter the equation (hence care-less). Data exhaust, as it is called, is imagined as a byproduct of living or interacting in the digital world – not worthy of care – and, yet, like the smoke from a tailpipe, is not inconsequential. The exhaust metaphor is particularly evocative in drawing attention to the materiality of data – that storing and circulating it, regardless of its designation as waste, requires huge amounts of energy and water (Marcus Law, 2022).¹³

The other narratives from my collaborative fieldwork illustrate two modes of care: one that works with a logic of transparency, and another that is driven by a constitutive relationality. Their juxtaposition with the first narrative illuminates how informationalization and datafication under racial surveillance capitalism necessarily require the creation of a so-called Global South, regardless of the specific place or location of this produced southernness (Trefzer et al., 2014).¹⁴ Like the first narrative, care grounded in a logic of transparency highlights how care-lessness is part of systemic design, and like the first narrative, it illustrates how the separation of people and data is also ultimately the severance of people from place, and their ultimate repositioning as marginal and vulnerable nodes in a global network of dependencies. The second kind of care evident in our fieldwork, grounded in constitutive relationality, offers a narrative of hope for those of us who use care as a way of revitalizing and revisualizing how to conceive of ways to cope with the severely damaged earth and her people – what Donald would call engagement at the frontier of difference. Among other lessons, it teaches us the wisdom of holism, the critical importance not separating people from place, and in the context of this essay, not separating people from data.¹⁵ Treating people, data and place as constitutively relational thus helps build ground for meaningful alternatives only to the care-lessness of systems of data extraction, but also to datafied notions of care. As environmental communication scholars, teachers and practitioners, it is particularly urgent that we write and teach the ethical and political urgency of enacting caring practices that center such constitutive relationships between people, place and data.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by Mistra, the Swedish Foundation for Strategic Environmental Research's Environmental Communication program.

Notes

1 Despite its presentation as such in this brief introduction, the idea that digital is material should not be reducted to a simplistic binary between physical and virtual. For an estimation of the emissions produced by digital exhaust, refer Widdicks et al. (2022).

2 For a comprehensive discussion and critique of Neel’s work, refer Radin (2017b).

3 Two early examples of the use of the term “digital exhaust” can be found in La Rosa et al. (2009), and in Krough (2009).

4 This is a point implied by the Care Collective (2020) and Pezzulo (2023).

5 For an overview of the work of this collective, refer to Kukutai and Taylor (2016).

6 For an analysis of this framework, refer Robinson et al. (2021).

7 The Care Manifesto.

8 This project was approved by the research ethics office at my university.

9 Andrew Barry (2004) refers to ethical capitalism in governmental terms: as a set of techniques and discursive practices that surface and ultimately comprise the ethical content of capitalist enterprises, rather than a prima facie claim that capitalism itself is more or less “ethical”.

10 As my colleague Jutta Haider observed in conversation, such benevolence is not only a manifestation of global white pastoral power, it undermines and prevents other forms of self-organization, including unionization.

11 My colleague Sharmila Rudrappa has written about this evocatively in her discussion about the relationship between garment work and surrogate motherhood in India.

12 Robin Wall Kimmerer’s influential book Braiding Sweetgrass vividly conceptualizes and illustrates this deep relationality.

13 The Standard Performance Evaluation Corporation for example calculated that server power consumtion in the US has increased 266% since 2017 and now takes up about 3% of global electricity consumption.

14 This stance is in keeping with scholarly work on shifting peripheries in the global north-south dialectic that suggests that one can exist within the other.

15 Refer Ramasubramanian and Dutta (2023) for a framework for data justice which implicitly inshrines this principle.