Autonomy in Relation

In the AI Story

The concept draws on Escobar's decades of engagement with Afro-Colombian communities in the Pacific region of Colombia, whose territorial governance arrangements had maintained community autonomy for generations despite continuous pressure from colonial, national, and global economic systems. These communities did not reject engagement with external systems. They engaged selectively, on terms shaped by their own governance structures and evaluated by their own criteria of well-being. The practice was real but precarious — constantly threatened by the structural adjustment programs, extractive industries, and development interventions that sought to absorb the territory into national and international markets on terms defined elsewhere.

Applied to AI, autonomy in relation suggests a different relationship to AI tools than either full adoption or full rejection. The Zapatista communities in Chiapas demonstrate the practice: they use mobile phones for communication while declining to participate in the digital platforms the phones enable access to. They have not rejected the technology. They have established terms — community-defined terms — under which the technology is deployed. The distinction between the capability and the platform is analytical, and the analysis is theirs. Similar practices exist among indigenous communities in the Amazon, agroecological farmers in West Africa, and many other communities navigating the gap between tools and terms.

The concept requires specific preconditions. It requires communities with sufficient internal coherence, governance capability, and analytical sophistication to evaluate external technologies against community-defined criteria. It requires institutional recognition — by states, by markets, by development agencies — that communities have the right to engage selectively rather than being absorbed into standardized frameworks. It requires the existence of alternatives — other communities, other tools, other networks — that make selective engagement practically possible. When these preconditions are absent or eroded, autonomy in relation becomes difficult to sustain, and communities are pushed toward the false choice between assimilation and isolation.

For AI governance, autonomy in relation implies institutional arrangements that the current political economy does not provide. Data sovereignty frameworks that recognize communal governance of knowledge used in training. Cooperative and multi-stakeholder governance structures for AI companies. Participatory design processes that include communities in foundational decisions. Regulatory requirements that protect community capacity to decline AI tools without losing access to essential services. Each of these represents a departure from the current concentration of authority, and each has precedents in other domains that demonstrate feasibility.

Origin

The concept emerged from Escobar's ethnographic work with the Proceso de Comunidades Negras and other Afro-Colombian and indigenous movements in the Pacific region of Colombia. It received theoretical articulation in Territories of Difference (2008) and was extended into the pluriversal framework in Designs for the Pluriverse (2018).

It draws on the broader Latin American tradition of thought on autonomy, particularly the Zapatista concept of autonomía and the indigenous movements' framework of buen vivir / sumak kawsay.

Key Ideas

Not isolation. Autonomy in relation does not require communities to separate from global systems but to engage with them on community-defined terms.

Selective engagement. Communities adopt capabilities that serve their purposes and decline those that do not, based on their own evaluative frameworks.

Preconditions matter. The practice requires internal coherence, governance capability, institutional recognition, and the existence of alternatives.

Community-defined criteria. The criteria by which technology is evaluated are set by the community, not by the providers of the technology.

Institutional implications. Autonomy in relation requires data sovereignty, governance pluralism, and participatory design as institutional infrastructure.