Illustration by Somnath Bhatt

Living with seamful technologies

A guest post by Ranjit Singh. Ranjit has a doctorate in Science and Technology Studies (STS) from Cornell University, and is a Postdoctoral Scholar at the AI on the Ground Initiative of Data & Society Research Institute.

This essay is part of our ongoing “AI Lexicon” project, a call for contributions to generate alternate narratives, positionalities, and understandings to the better known and widely circulated ways of talking about AI.

\ ˌim-brə-ˈkā-shən \

Definition:

  1. an overlapping of edges (as of tiles or scales)

The tribal areas of Rajasthan [a state in the west of India] are very hilly. How would you get network [internet] there? Let me give you an example. The Fair Price Shop is in the middle of one of these villages in the hills. The ration [subsidized food grains] is there; the shop owner is there; the POS machine [Point-of-Sale machine equipped with fingerprint readers] is there; and so, the villagers went there. When the villagers put their thumbs on this machine, it did not work. There was no network. The shop owner decided that the network problems were happening because they are in a valley, so he decided to hike up a hill to look for network. The villagers followed him around with their ration cards to find network on top of the hill. Once they found network, they put their thumbs on the machine again.

[At this moment, the narrator was interrupted with a question: What about the elderly and the disabled?]

The elderly and the disabled, of course, cannot walk up a hill. They remain excluded. By chance, if they are successful with [Aadhaar] authentication, they go down to the shop later and collect their ration (A Right to Food activist narrating troubles of beneficiaries with Aadhaar authentication in accessing subsidized foodgrains, 26 July 2016).

Aadhaar (translation: foundation) is India’s biometrics-based national identification infrastructure built to streamline government services and reshape the nature of state-citizen relations. Upon enrollment, every Indian resident is given a unique 12-digit Aadhaar number based on their biometric (fingerprints, iris scans, and facial photograph) and demographic (name, age, gender, and address) data. In my dissertation research, I focused on the limits and uses of Aadhaar to reorganize the Public Distribution System (PDS) of subsidized foodgrains to below poverty line families in India. The above excerpt is from a meeting of Right to Food (RTF) activists gathered from all different states in Delhi. They were given the stage for about 15 to 20 minutes to narrate the challenges of securing welfare in their respective states. The RTF activist from Rajasthan took his opportunity to speak at length about struggles with Aadhaar authentication to access PDS entitlements. Beneficiaries literally had to walk up a hill to secure their welfare benefits.

This excerpt is not simply a story of digital divide, although it can certainly be analyzed as such. I use this story to showcase that government services cannot be digitized in a day. It is a long process that spans years and requires a deeper engagement with how new digital technologies interface with existing practices of governance. For example, consider the question of attributing accountability for the challenges faced by the above PDS beneficiaries. Who is at fault here? Is it Aadhaar? Is it the network provider? Is it the POS machine? Is it the policy of using Aadhaar in determining access to PDS entitlements?

The answers are neither straightforward nor obvious. Yet in arguments for or against digitally (re)organizing services and associated datafication of individuals writ large, there is a totalizing narrative at play. Consider common arguments such as “services will become more efficient with the use of data” or “data is a tool to profile and surveil individuals.” The underlying work practices that make these arguments possible as emerging conditions of everyday life are either assumed to work seamlessly or made invisible. These processes, however, are not seamless; digital technologies do not neatly interface with and layer over existing practices. Instead, they are seamful and imbricate with existing practices. This essay explores ‘imbrication’ as a metaphor and an analytic resource to engage with master narratives of ‘seamlessness’ with which most technological systems are presented to the world and their users. As governments promise progress and digital inclusion, I argue that we should pay attention to the seamfulness of digital systems, the gaps in practices of leveraging them to deliver services, who they leave out, and how they overlap with our daily lives.

I initially encountered “imbrication” in a volume edited by Lampland and Star, where it was used in opposition to the ‘stack’ metaphor that is often used to signify seamless layering of components on top of each other in computing practices. Challenging this image, Lampland and Star used the metaphor of a stone fence to illustrate how digital technologies, their components, and elements of existing practices overlap unevenly to produce a sociotechnical assemblage. This imbrication produces the conditions of infrastructuring data into everyday life and by extension, the lived experience of datafication. Imbrication signifies that layering of digital technologies with existing practices is imperfect, incomplete, and inevitably seamful. Stories of both data-driven innovation and marginalization emerge in contending with the seamfulness of these systems and the gaps between them and practices of leveraging them. When applied to the context of using Aadhaar to distribute PDS entitlements, this insight allowed me to move beyond the idea of biometrics as a new digital technology to address diverse problems of Indian governance. I began investigating the changes in practices of governance that preceded and accompanied Aadhaar as its imbrication with these practices became the “foundation” of state-citizen relations in India. In following this line of inquiry, I started researching ongoing forms of work-at-the-seams to use digital technologies for governance. How did biometrics become a condition to access PDS entitlements? What do beneficiaries and last mile service providers do when digital systems fail? The starting anecdote of this essay illustrates this work-at-the-seams.

While imbrication is often used across academic disciplines by researchers to engage with the politics of uneven overlaps between their phenomena of interest, it has profound consequences for studies of sociomateriality of digital systems (especially, organizational studies that showcase how social and material conditions are inextricably intertwined in building and appropriating digital technologies). Paul Leonardi has enumerated a set of “methodological marching orders” to study how digital systems become a part of our everyday material lives: “explain how and why imbrication occurs, why certain practices come to take on the shape they do, and why people think those practices had to occur as they did” (Page 71). To study imbrication is to explore how our experience of being subject to datafication is enabled by making do with digital systems that partially overlap with each other. This making do, however, also produces challenges to our lived experiences with data. Imbrication works for some at the expense of others; it produces difference in everyday lives of people increasingly made subject to a data infrastructure-in-use. For example, many PDS beneficiaries believe that their access to their welfare entitlements has become more reliable through Aadhaar-based authentication. It has also raised significant barriers and challenges for others who were able to access their welfare entitlements previously based on a paper-based identification procedure. Some people fall through the cracks (such as the elderly and the disabled in the starting anecdote); others must navigate the gaps (such as the beneficiaries who walked up a hill to find ‘network’). Study the imbrication, thus, orients us to describe not only how systems work but also how they break at their seams. When systems break, people are doubly challenged: (1) they experience and contend with failures in accessing digital services; and (2) they contend with conditions of diffused accountability for such failures such that fault can be located at either side of these seams.

Given the fractured data environments in the Global South where the current focus is on finding ways to digitize services, a call to study the imbrication is an invitation to decenter data and digital technologies in discourses of modernity, progress, and infrastructural development. Instead, focus must turn to antecedent and emerging work practices that together produce the conditions for using data and digital technologies in the pursuit of intended goals (such as infrastructural development or accountability in last mile delivery of social welfare). This method is as much sociomaterial as it is political. It also requires analysts to account for choices they make to focus on certain practices as opposed to others in the labyrinth of standardization that makes up practices of infrastructuring data. Such standardization is accomplished through a wide variety of interventions in different domains of state practices, such as technological design, administrative procedures, policy work, and court cases. Studying any of these interventions presents distinct sets of challenges and requires different analytic skill sets. Since standardization of data practices is increasingly becoming central to organizing for the state, it is imperative to make choices about which practices to focus on and how. As Susan Leigh Star once put it:

We live in a world where the battles and dramas between the formal and informal, the ill structured and the well structured, the standardized and the wild, are being continuously fought. These battles are sometimes benign and sometimes tremendously helpful to humankind, such as the standardization of climate change data (or attempts to do so). However, attempts to overstandardize (using tools such as electronic surveillance) are haunting social justice. So thickly imbricated are these battles now with electronic life and daily offline life that it is no longer a question of choice. If not now, when? (Page 614–615)

She reminds us that we never begin “from scratch” when infrastructuring data. New practices do not have a singular “foundation;” they draw on and extend a variety of existing practices. After all, imbrication is never done once and for all.