Digital Colonialism Is an IS Theory Problem, Not Just Politics

I was reading Bardhan et al.'s 2025 review in Information Systems Research and stopped cold at the phrase "data colonialism." It appeared in a paragraph about scaling technology across diverse healthcare systems, wedged between data poverty and the need for population-specific synthetic data. The authors were not writing a political manifesto. They were mapping future IS research clusters. Still, the phrase carried weight. I flipped back through my notes and saw institutional theory sitting a few pages earlier, with its neat list of coercive, mimetic, and normative pressures. The two ideas did not belong together in my head until that moment. Then I realized they describe the same pattern at different scales.

When a government in a developing country adopts Amazon Web Services for its national cloud, or requires civil servants to use Microsoft 365, or trains its engineers on TensorFlow, we usually explain this through efficiency. Better tools, lower cost, faster deployment. DiMaggio and Powell (1983) would disagree. Institutional theory predicts that organizations become similar not because convergence is optimal, but because three pressures make deviation costly or even unthinkable. Coercive pressure comes from external regulation, and in the global technology market it includes trade rules, sanctions, and the simple absence of viable alternatives. When your banking system depends on SWIFT and your mobile ecosystem on Android, the pressure is structural, not optional. Mimetic pressure kicks in under uncertainty. If you are a technology minister deciding whether to fund a local search engine while Google serves ninety percent of your population, imitation is the rational choice even when the long-term cost is dependence. Normative pressure flows through professional training, certification programs, and the global circulation of engineers who learned their craft on Silicon Valley stacks. The result is institutional isomorphism on a planetary scale. The platforms look the same everywhere because the field has been flattened by legitimacy, not by merit.

Scott (1995) extended this logic with a cultural-cognitive pillar that makes the pattern harder to reverse. When a technology becomes taken for granted, the idea of building a sovereign alternative stops looking ambitious and starts looking naive. I saw this firsthand in conversations with startup founders in Tehran years ago. Their objection to building local platforms was never technical. It was that users refused to leave Instagram and Telegram because those apps were simply what messaging and photo sharing meant. The technology had become the ontology of the practice. That is the cultural-cognitive pillar operating in real time. Once a platform defines the category, competing with it is not a business problem. It is a category error.

Resource dependence theory explains why the alternatives stay weak. Pfeffer and Salancik (1978) argued that organizations survive by managing their dependence on critical external resources. Wade and Hulland (2004) drew on this tradition to show how environmental munificence shapes which IS assets create value. In the global technology market, munificence is asymmetric. Venture capital, cloud credits, developer documentation, and user networks concentrate in the United States and diffuse outward. A local platform startup faces a resource environment that is thin on every dimension that matters. There is no equivalent of Y Combinator in most developing markets, no comparable base of paying enterprise customers, and no cross-border network effects to match those of incumbent platforms. Schryen (2015) noted this explicitly in a review of IS literature, observing that interorganizational information systems often function as instruments of dependence. The platform owner controls the resource; the complementor adapts or exits.

Recent reports suggest that several governments have begun pushing back. India has moved toward sovereign cloud requirements. Brazil has debated data-localization rules. Russia has long maintained its own social network ecosystem, and China built the most complete parallel platform stack in existence. These measures are usually framed as sovereignty or security. From an IS theory perspective, they are attempts to break resource dependence and resist isomorphic pressure. They are also enormously expensive. Building a domestic cloud or a locally hosted large language model requires capital, talent, and time that could be spent adopting the ready-made American alternative. The cost of autonomy is the cost of illegitimacy in a field where the dominant players have already defined what counts as normal.

Mayer et al. (2025) studied how generative AI reshapes digital platform governance by acting as a boundary resource. Boundary resources are the interfaces, tools, and rules that platform owners use to coordinate complementors. The authors found that complementors resist and accommodate over time, reshaping governance through conflict. The global version of that dynamic is what we are watching now. National governments are complementors who suddenly find the boundary resource too open-ended to trust. Their resistance is not merely political defiance. It is a governance negotiation at the scale of international relations. But unlike corporate complementors, states have fewer exit options. They cannot simply move to another platform ecosystem unless they build one from scratch, which brings us back to resource dependence and the steep penalty for deviance.

Bardhan et al. (2025) cite Arora et al. (2023a) for the specific claim that algorithmic bias and data colonialism marginalize populations by exporting data infrastructure designed for wealthy markets. The argument is not that technology is inherently harmful. It is that the institutional and resource structures surrounding it create systematic asymmetries. Local health data trains models in Palo Alto. Local content creators earn revenue through platforms headquartered in Menlo Park. The data flows out; the infrastructure decisions flow in. This is not a conspiracy. It is the predictable outcome of institutional isomorphism combined with resource concentration.

I keep returning to the question of what IS researchers are supposed to do with this. We have frameworks that explain conformity and dependence beautifully. We are less practiced at using them to explain resistance and alternative design. Institutional theory predicts convergence. It says less about deliberate divergence. Resource dependence explains survival strategies under constraint. It does not tell us what a sovereign stack would look like if the constraint were removed. Arora et al. (2023a) opened the door by naming data colonialism as an IS problem. The next step is to treat platform imperialism as a theory-testing moment. When a country blocks a Western platform and builds its own, it is not just politics. It is an experiment in breaking isomorphism. Someone should study it.