Philosophy
What if data taken from marginalised communities had to pass through something like a gram sabha before it could be used?
Data Sovereignty and the Grammar of Consent
The gram sabha, the village assembly guaranteed under the Forest Rights Act of 2006, is not merely an administrative body. It is a philosophical institution. It embodies the principle that the communities most intimately connected to a place have the prior right to decide what happens to it. Before a mine can be dug, before a road can be cut through, the gram sabha must consent. This consent is not a formality to be managed; it is a structural veto, the recognition that no external calculation of value can override the judgment of those who live in relationship with what is being valued. Now consider how data is collected from marginalised communities in India. Consider the health records of Adivasi patients, the agricultural patterns of small farmers enrolled in government schemes, the movement data of urban migrant workers, and the social networks of Muslim women in minority-concentration districts. This data is gathered, processed and fed into AI systems for credit scoring, for welfare targeting, for predictive policing, for medical diagnostics, without anything resembling Gram Sabha-style consent. There is no village assembly. There is, at best, a checkbox.
This is not an accident of implementation that better policy could fix. It is a structural feature of how data is understood. Data, in the dominant framework, is a byproduct, a residue of activity that the person engaging in that activity does not fully own, because ownership requires a particular kind of legal personhood, a particular kind of claim. Communities that have been systematically excluded from legal personhood, such as Adivasi communities, Dalit communities, informal-economy workers (to name a few), find that their data is among the most valuable things they produce and the thing over which they have the least control. Indigenous data sovereignty frameworks developed elsewhere offer a different grammar. The CARE Principles: Collective Benefit, Authority to Control, Responsibility, Ethics, developed by the Global Indigenous Data Alliance, insist that data governance cannot be reduced to individual privacy rights. Data about a community belongs, in a meaningful sense, to that community. The Māori data sovereignty movement in Aotearoa New Zealand has gone furthest in institutionalising this: Te Mana Raraunga, the Māori Data Sovereignty Network, argues that data is not merely information, it is an expression of whakapapa, of relational identity, and must be governed accordingly.
India already has the conceptual and legal infrastructure for an equivalent framework. The Forest Rights Act established, however imperfectly, that Adivasi communities have collective rights over their forests. The gram sabha’s prior informed consent is a legal principle with philosophical depth. What would it take to extend this logic to data to insist that communities have collective authority over the data generated by and about them? genuine veto power over how it is used, and a meaningful share of whatever value it generates? This is not a utopian demand. It is the application of an existing principle to a new domain of extraction.