Is data colonialism a universally defined legal category?

This entry uses it as a critical social-theory concept, not a legal determination.

Is the argument only about foreign companies collecting data?

No. The authors' account also concerns relations within societies.

Can record counts prove the label applies?

No. The argument concerns social arrangements and requires evidence about them.

AI Glossary / Data engineering and architecture / Advanced

Data colonialism

Definition

Data colonialism is a critical social-theory concept used to examine the appropriation of human life and activity as data for economic power. In Nick Couldry and Ulises Mejias's account, it connects contemporary data extraction with structures of appropriation associated with colonialism, while recognizing differences from historical territorial rule. It is an analytical argument about social relations, not a technical data format or a numerical classification test.

What it means

The concept asks who can turn people's activities into valuable data, who controls the resulting systems and how the benefits and burdens are distributed. Its focus extends beyond whether a particular file contains personal identifiers. A discussion using this framework might examine dependence on a platform, the ability to refuse participation and the terms under which collected information becomes a commercial asset. Readers should distinguish the authors' theoretical interpretation from independently established facts about any particular organization or project.

How it works

Couldry and Mejias describe data relations that organize everyday activity so it becomes available for extraction and commodification. Their argument is broader than an allegation that one organization collected too much information. Applying the framework involves examining the arrangements around collection, control and value, then making the reasoning explicit. The label alone supplies no evidence. A careful analysis identifies the actual practices under discussion and explains why the theory illuminates them, without assuming that all forms of data use have identical purposes, power relations or consequences.

Three practical examples

A seminar uses the concept to examine who controls information produced through a widely used digital service.
A project review asks whether a community contributing data has meaningful influence over later uses and benefits.
An AI-policy discussion compares the provider's account of a data partnership with the experiences and choices of people represented in the dataset.

Think of it this way

Imagine a shared garden whose visitors generate observations about paths, habits and preferences. One organization records those observations and builds a valuable service from them. Asking how many observations were collected describes the volume. Asking who set the terms, who can challenge the use and who benefits examines the relationship. This is an original teaching analogy for questions raised by the concept, not an assertion that the garden arrangement reproduces historical colonial rule in every respect.

Worked example

Consider an original fictional comparison involving two projects, each collecting 10,000 observations from a community. Project A gives contributors a role in defining purposes and reviewing proposed uses. Project B gives contributors no such role in the stipulated scenario. The observation count is identical, so counting records cannot distinguish those governance arrangements. That difference provides material for a discussion of power and control, but it is not a validated formula for declaring one project colonial and the other non-colonial. Additional evidence about dependence, alternatives, benefits and actual practice would be needed for a substantive argument. The example demonstrates a question to investigate, not a diagnostic score.

Illustrative case study

A fictional research partnership proposes collecting local language recordings to develop an AI service. Participants welcome the proposed service but ask who will control later uses and whether they can influence decisions once collection ends. The partners revise their proposal to make those arrangements explicit and invite discussion of the expected benefits. A scholar might use data-colonialism theory to analyze the original and revised relationships. The scenario illustrates how the framework can direct attention to institutional choices. It does not establish that every language-data project is exploitative or that one governance change resolves every concern.

Common misconceptions and mistakes

The term should not be treated as a synonym for any database, international data transfer or commercial use of information. Nor does using it establish that contemporary practices are identical to historical colonial violence or territorial occupation. Another mistake is reducing the argument to a count of collected records. A small collection can raise important power questions, while volume alone does not settle an interpretation. Finally, presenting the theory as an undisputed technical fact obscures its analytical character; claims about a specific case need evidence and clearly stated reasoning.

Business use

The concept can broaden a discussion of an AI or data initiative beyond its immediate technical performance. Examine whose activities supply the information, who determines subsequent uses and how affected people can influence decisions. Ask whether the stated benefits match the actual distribution of value and responsibility. These questions can reveal assumptions that a purely operational review overlooks. Use the framework to structure inquiry rather than to replace evidence with a label. A useful business discussion should identify concrete arrangements that could be clarified or changed and explain why they matter to participants.

Check your understanding

Why did the two toy projects have the same observation count?

Answer: To show that volume alone did not describe their different decision arrangements.

What did the example establish?

Answer: A distinction worth investigating.

What did it not establish?

Answer: A definitive classification of either project under a validated technical test.

Related terms

Datafication
Data governance
platform power
data sovereignty
commodification
community participation

Learn next

Choose a hypothetical data partnership and map the participants, decisions and expected benefits. Separate factual descriptions from interpretations. Then explain which aspects a data-colonialism analysis would question and what additional evidence would be needed to support or challenge that interpretation without relying on the label alone.

Sources

Nick Couldry and Ulises Mejias, Data Colonialism: Rethinking Big Data's Relation to the Contemporary Subject, accepted manuscript: https://eprints.lse.ac.uk/89511/1/Couldry_Data-colonialism_Accepted.pdf The two-project comparison, garden analogy and language-recording case are original educational illustrations. The article is attributed as a theoretical argument; the examples are not findings about named organizations or a legal assessment of any actual project.

Previous: Data collection Next: Data complexity