Co-Production of Knowledge and Social Order

In the AI Story

Jasanoff developed co-production through her studies of regulatory science — the knowledge produced at the intersection of scientific inquiry and institutional decision-making. In The Fifth Branch, she demonstrated that regulatory agencies do not simply 'apply' scientific knowledge to policy problems. They produce a hybrid form of knowledge shaped simultaneously by scientific methods and by the legal, political, and institutional contexts in which the science is conducted. A drug is not safe or unsafe in the abstract; it is safe or unsafe within a specific regulatory framework that defines what counts as evidence of safety, who produces that evidence, and what standard of proof is required. Change the framework and you change the knowledge — not arbitrarily, but consequentially.

The concept found its fullest development in States of Knowledge (2004), where Jasanoff compared how the United States, the United Kingdom, and Germany produced authoritative knowledge about biotechnology. The three nations faced the same scientific questions but arrived at different answers because they had different institutional practices for producing and validating knowledge. These practices were not neutral instruments for discovering truth; they were co-productive of the truths they discovered. The knowledge about genetically modified organisms that emerged from American adversarial hearings was genuinely different from the knowledge that emerged from German consensual expert panels — different not because one was right and the other wrong, but because knowledge is shaped by the process of its production.

Applied to AI, the co-production framework reveals that when Edo Segal's team in Trivandrum experienced a twenty-fold productivity gain, what was being co-produced was not merely more code but a new social order within that engineering organization. The hierarchy of skills shifted — backend knowledge became less scarce, cross-functional capability became more valuable. The distribution of authority changed — individual engineers gained power relative to teams, judgment gained value relative to execution. Professional identities were reorganized — the specialist became a generalist, the implementer became a director. These social transformations were not consequences of the technical change; they were the technical change, viewed from a different angle. The code and the org chart were being written simultaneously.

The governance implications are profound. If the technical and the social are co-produced, then governing AI cannot mean governing the technology alone. It means governing the social order that the technology produces — the new hierarchies, the new distributions of power, the new definitions of expertise. And governing a social order requires democratic participation. The people whose social world is being remade must have voice in the remaking — not as beneficiaries of decisions made by others, but as participants in the decisions themselves. The current AI governance landscape treats the technology and its social consequences as separable, addressing each through different institutional channels. Co-production reveals this separation as a category error that guarantees governance failure.

Origin

The term 'co-production' has multiple intellectual genealogies. Jasanoff's usage is distinct from the public administration concept (citizens co-producing public services) and the economics concept (joint production by complementary inputs). Her version emerges from science and technology studies, particularly from the Edinburgh Strong Programme's principle of symmetry — that true and false beliefs should be explained by the same sociological causes — and from Bruno Latour's actor-network theory, which dissolved the boundary between the social and the technical. Jasanoff's distinctive contribution was to apply this dissolution not merely to the analysis of science but to the governance of technology, showing that the co-production of knowledge and order is not merely an analytical observation but a practical problem that governance institutions must address.

Key Ideas

Technical and social changes are simultaneous. The standard narrative treats technology as developing first and society as responding afterward. Co-production reveals that the technical capability and the social reorganization are produced in the same moment by the same process.

Knowledge is shaped by institutional context. What a society knows about a technology depends on the institutions through which that knowledge is produced — the evidentiary standards they apply, the voices they include, the questions they recognize as relevant.

Separation is a governance failure. Governance frameworks that address 'the technology' through technical regulation and 'its social effects' through separate policy instruments have already failed, because they treat as separable what is actually inseparable.

Democratic participation is not optional. Because technology and social order are co-produced, governing technology requires governing the social order, and governing social order requires the participation of the people whose order is being made.