Information infrastructure and democracy

2015-07-15

As the ‘open’ movement builds momentum, information practices are being tied to a range of aspirations: reproducible science, sustainable development, and accountable government, to name a few. Many advocacy and technical efforts across open [knowledge, data, science] focus on revealing, unlocking, and ‘openly’ digitally publishing information that currently circulates—or sits—in restricted, enclosed, or encumbered systems. But the movement is multi-stranded, and there are undercurrents of aspiring to even deeper change—perhaps radical changes in the relationship of knowledge and society. This is what excites me the most.

So, I was pleased to read a recent article by Jonathan Gray about the politics of public information in the “data revolution.” [1] The issue that Gray raises is this: while we may push for transparency of public data, that won’t by itself ensure democratic, evidence-based collective action. Without democratic governance of public information infrastructures—for example, deciding what is and isn’t measured in data collection—and without democratic initiatives to use these for progressive change, the data revolution may be ineffectual, or even harbour social and economic risks. Gray has elsewhere used the term “participatory data infrastructures” to describe what’s needed for a more democratic politics of information. [2]

The idea that information infrastructures should be democratized is worth repeating and elaborating.

The infrastructure

First of all, to take this useful concept from Geoffrey Bowker and Susan Leigh Star’s book Sorting Things Out, [3] “information infrastructures” are ubiquitous, embedded, far-reaching and long-lasting networks of technological and informational tools. Standardized classification systems are the primary examples that Bowker & Star explore in their book, but everything from the metric system to QWERTY keyboards could be considered as infrastructure—and certainly things like internet technologies and the standards used in government data collection (census forms, etc.).

Why does this matter? Because information infrastructures are important elements of the social order. They both enable and constrain us in what we can know, do, or even be. More broadly, we can draw on the work of democracy theorist Richard Sclove, who regards technology as a form of social structure. [4] Like law and policy, technologies shape the basic conditions of our lives: by structuring immediate social relations, conditioning patterns of work or community life, creating opportunities and constraints for people’s action and behavior, helping to configure power relations, and even by contributing to how we define ourselves as people.

All of this applies to information infrastructure and even more so, since we use these systems to encode knowledge—including data about people—and also to represent people themselves. We all appear in administrative, economic, and medical statistics. ‘Social data’ represent us as consumers of information and of products; by analyzing and nudging our behavior, systems like Facebook and Amazon seem to make us converge with that representation.

Interestingly, Sclove twenty years ago foreshadowed some of this current discussion. He wrote parenthetically: “While there is some hope for emerging international information systems to facilitate transnational grassroots political deliberation and coordination, it seems obvious that for the time being the world’s affluent professionals will be disproportionately active in global electronic communications networks” (p 237). It would have been more accurate to say “disproportionately active in shaping global electronic communication networks.” He also foreshadowed “invasive electronic surveillance of [consumer] shopping habits and political predilections and … unwanted medical tests and genetic discrimination” (p 232). Let’s hope that last one still fails.

Why democratize?

Infrastructure’s role in shaping our reality generally recedes into the background. Since it’s meant to provide already-installed tools or capacities that we don’t need to constantly reinvent (like electricity or hypertext transfer), we easily forget about it. At the same time, it becomes naturalized: we see it as the way things are, and always have been. The more additional systems come to depend on a piece of infrastructure (i.e. compatibility), the more indisputable it becomes.

Yet, technologies and information infrastructures have politics. Ideologies and values are embedded in them, even if they’ve been erased from view as the system has naturalized. And what if we’d like to question those ideologies? Bowker & Star say this about classification systems:

One of this book’s central arguments is that classification systems are often sites of political and social struggles, but that these sites are difficult to approach. Politically and socially charged agendas are often first presented as purely technical and they are difficult even to see. As layers of classification system become enfolded into a working infrastructure, the original political intervention becomes more and more firmly entrenched. In many cases, this leads to a naturalization of the political category, through a process of convergence. It becomes taken for granted. (p 196)

You have to work really hard to dispute the ideas built into the infrastructure. It took activism to remove homosexuality as a category from the Diagnostic and Statistical Manual of Mental Disorders (DSM); it took activism to allow U.S. citizens to select multiple racial categories on the census. This was political activism directed at the infrastructural conditions for the production of data and social order.

The problem requiring democratization is that we have no established institutions that are adequate for democratically deliberating on matters of technology (as Sclove argues)—and the same goes for information infrastructure.

The idea of participatory data infrastructure design seems like a great place to start. It could complement or build on two approaches that have already been developed: participatory technology assessment [5] and participatory design. [6]