Searching for Parables of AI
Ranjit Singh and Rigoberto Lara Guzmán discuss the evolution of their project and the relationship of stories to research.
In October 2021, Data & Society researcher Ranjit Singh and senior producer Rigoberto Lara Guzmán hosted an experimental workshop that was designed to produce new ways of engaging with the global impacts of digital technologies, with an emphasis on storytelling as both a research craft and an intellectual practice. Yesterday, we released Parables of AI in/from the Majority World, an anthology (edited and curated by Singh, Guzmán, and Patrick Davison) that grew out of that gathering. It brings together twelve original stories about the everyday experiences of living with AI-based systems from storytellers in Africa, Latin America, the Middle East, and South Asia. The process that led to this collection was unconventional and highly collaborative: Here, Guzmán and Singh reflect on the project’s evolution, the importance of stories to research, and the far-reaching potential of parables.
The collaboration that sparked this project is important to its final form. Can you talk about the nature of that collaboration and what you each learned from it?
Singh: The first time I met Rigo was for my job interview. I was late; Rigo had already been waiting for me for about 15 minutes by the time I got on the call. Yet, we got past that initial hiccup and had an amazing conversation — one that became even more crucial for me when I joined D&S. We were all remote, and Rigo became one of my go-to folks; I learned to be part of D&S through Rigo. Then in early 2021, we worked together to plan the workshop on AI in/from the Global South. My focus has always been on research, and turning a research endeavor into a community gathering was all thanks to Rigo’s incredible experience in producing events. The anthology is the way it is because the workshop ended up shaping the submissions we got, the storytellers we engaged with, and the community that we built through it. So the sense of relationality, and the relationships we built in producing this anthology, come from a place of deep generosity in our collaboration.
Guzmán: Madeleine Clare Eilish was the director of the AI on the Ground initiative at the time, and her decision to invite me into the hiring process was indicative of a turning point at D&S. I was not part of the research team in any significant way, yet she made this gesture of opening up a space. I took that seriously, because there have typically been silos between research and non-research staff. Being invited to contribute to hiring a post-doc position interrupted the silo and set the stage for what would become the core ethics of my working relationship with Ranjit. During the interview, I asked him questions like: Who are your people? Who are you accountable to in this world? How do you define a network and how do you enact network maintenance? The values inherent in these questions have persisted in our collaboration; they became key principles for organizing the workshop and later the anthology. They combine the two pillars that are so characteristic of our work at Data & Society: empirical research and network building. This collaboration has been so fruitful because we found a way to merge both.
You didn’t initially set out to do a storytelling project. It grew out of a different kind of inquiry into mapping AI in/from the global south or majority world, and specifically digital IDs. Now that the project has taken on the form of this anthology, how do you think about the relationship between the work of mapping and storytelling?
Singh: One of the orienting stories for my dissertation research (on India’s biometrics-based national identification number, Aadhaar) came from a conversation with one of my earliest field interlocutors. She narrated her troubles in accessing government services without having an Aadhaar number. As she was describing what was nothing short of a Kafkaesque experience, she said, “computers are the new bureaucrats.” That laid the groundwork for me to think about how Indian citizenship is increasingly mediated by computers. The most interesting part of the initial interviews I did with practitioners and experts for my project on mapping AI in the global south were their stories from the field, which often gave me an insight into why their work or research took the direction it did. The work of mapping the field increasingly became about storytelling, because the most persuasive parts of my research were stories from the field. How do we tell stories of the research we do? My initial idea was to focus on dispatches or stories from the field. Rigo convinced me to drop the “field.” Practitioners have stories. Researchers have stories. Everybody has a story, if you want to listen to it.
What does the idea of a parable mean to you, and why is it meaningful to this project?
Guzmán: In a broad sense, we’re interested in how concepts are generated or conceived. How do narratives begin to gain traction in a certain kind of contextually bound discursive exchange? We have been talking a lot about shifting narrative power. What does that mean? What does it look like in practice? When do I know that a narrative lens has shifted?
“Parable” is a word emerging from an ecclesiastical Latin origin meaning “discourse or allegory.” In a mathematical sense, we can think about it as a parabolic curve; this visual of curving a discourse by placing it side-by-side with another begins to illustrate our methodology. A parable is also a simple story, used to illustrate a lesson: you might think of the Gospels in the Judeo-Christian tradition. While we recognize that lineage, we are drawing our sense of the word more from a Science and Technology Studies (STS) perspective, and thinking about the role of a story in illustrating an idea or a concept.
Singh: One of the most well-known parables in STS is Langdon Winner’s story about the design of low-hanging bridges over the parkways on Long Island, New York. They were built that way so that cars could pass through them, and buses could not. This design prevented the large number of inner-city residents (mostly poor and Black) who depended on public transport and did not own cars from getting to the beaches. The point of this story is that designing technology is a moral choice; these bridges perpetuate a racist form of order. There are long-term distributional consequences to the ways technologies are designed. While bridges may recede into the everyday background of New York City life, once you see these choices, you can’t unsee them. Another example is the spikes or metal bars on public benches. That is a design decision made to prevent people, especially homeless people, from sleeping there. These examples and stories not only illustrate a moral point but provide a way for listeners to take the story and its point with them, and share it with somebody else. What turns a story into a parable is the exercise of listening and sharing.
Guzmán: An illustrative parable allows a researcher to legitimate a claim or an argument. It is a citational practice; cite someone who has already made that claim and move on to building on the claim. These are norms and practices of knowledge production that we take for granted but, at the same time, they are indicative of the ethics that we imbue our concepts with. In a sense, the so-called “real world” is constructed out of concepts; it is imbued with value-based ideas that we often take for granted. In order for that bench with spikes to exist, it first had to be imagined. That is also why we spend so much time debating this word versus that one at D&S: as a para-academic organization, we inhabit a conceptual realm where ideas take root and then materialize out in the world.
I’m not going to ask you to do the impossible task of choosing a favorite story in the anthology, but is there one that has particular resonance for you and your work?
Singh: For me, it is Sacha Robehmed’s story about seeing Facebook ads for emigrating to Canada and Europe while she was living in Lebanon in the midst of the Syrian refugee crisis. She explores the various layers of meaning in these ads, starting with what receiving a “personalized” ad meant for her as an individual, what it indicated about her network of friends and her activities on social media, and finally, what the ad represents in the context of an ongoing Syrian refugee crisis in Lebanon. Her story is a really good example of what it actually means to encounter algorithms in everyday life. But it also goes further, by asking you to take a step back and think about what these ads mean and why they appear on your feed at a particular moment in time. There is this moment in Kimberly Fernandes’s story, where a father is reading aloud and explaining a bureaucratic form to his son. It is mundane and yet, so poignant. Similarly, when someone asks Andrea Pollio, “What about China?” it triggers a whole trajectory of fieldwork and research for him. I like these moments because they transform everyday experiences into occasions of deep reflection and sometimes even change.
Guzmán: For me, it would be Zehra Hashmi’s story about navigating the border between Pakistan and Afghanistan, and its implications for Pakistan’s biometrics-based ID. Identification is typically framed as being you as an individual: you get your personal number; you get your personal card; you have to have your own photo. But in her story it becomes all about the kinship relations you have and how you can sustain them bureaucratically. It also says a lot about how borders shape identification.
How do you see this work being connected to the Primer on AI in/from the Majority World, which D&S published earlier this fall?
Singh: Both publications are the result of the community of scholars that came together during the storytelling workshop. You can read the parables work through the lens of the primer, and vice versa. You can ground the primer in the stories from the anthology and you can read any of the stories through the conceptual vocabulary that the primary is trying to build. The work we’re trying to do is in between the two.
Guzmán: Both publications foreground forms of situated knowledge. In the primer, where we treat the majority world as an empirical site and a standpoint, that’s very deliberate. So often the notion of a standpoint is geared towards positionalities that are marginalized, or excluded, or undervalued, or undermined in some way. The whole point of standpoint theory is to be able to locate those as sites of expertise and knowledge, right? But lately, I’ve been contemplating the inverse when it comes to locating ourselves. How do we look at ourselves as having a particular kind of standpoint? I would venture to say that this work is really a result of what we might begin to describe as a kind of diasporic epistemology. Both publications foreground scholars working in/from diasporic communities outside of their place of origin. A great example here is Sareeta Amrute’s idea of “in/from”: the slash in between speaks so much about where this work is coming from.
The anthology’s epilogue outlines how others can organize their own storytelling workshops. That’s partly a way to reveal your own process, which is very valuable on its own. But I wonder about your deeper hopes and ambitions for others to pick up this model and run with it. What might that look like?
Guzmán: The overarching takeaway of both the epilogue and the anthology as a whole is that we want this process to be reproducible. We want it to be a proof of concept and a model that other people can take on, update, and implement. That would be a step toward realizing the core mantra of our work, which one of our partners, Dibyaduti Roy, phrased as seeding a “rhizomatic praxis in motion, like ginger.” He said this out loud during a planning call, and it just stuck. So you can think of the epilogue as our methods section: It lays out our process and partially reveals how the workshop came together. It is a gesture towards open source, procedural transparency.
