John MoravecNew paper: AI, coloniality, and educational sovereignty
The rapid adoption of generative AI in education has often been framed through the language of “readiness.” Institutions are encouraged to develop policies, train staff, clarify acceptable use, and prepare learners for participation in an AI-mediated future. These are important tasks, but they fail to address the elephant in the room. Readiness frameworks can obscure a more basic governance question: who holds authority over the conditions under which teaching, learning, assessment, and inquiry now occur?
My recent article in Learning Futures and Emerging Technologies, “Beyond the Digital Enclosure: AI, Coloniality, and Educational Sovereignty,” examines this question by treating AI adoption not only as a pedagogical issue, but as a problem of institutional authority and governance. The article argues that closed-weight, cloud-hosted, vendor-controlled AI systems can shift decision-making power away from public educational institutions and toward private platform firms. This shift occurs through technical infrastructures, procurement arrangements, data practices, classification systems, content policies, and interface design.
To analyze this process, I propose “four lanes of coloniality”: infrastructure, classification, epistemology, and labor. These lanes describe how platform power enters educational systems through vendor dependency, automated surveillance, linguistic bias, epistemic gatekeeping, and uncompensated repair work performed by educators and students. The framework draws on decolonial theory, digital enclosure, platform studies, surveillance capitalism, epistemic justice, and educational technology research.
| Lane | Primary mechanism | Pedagogical impact | Colonial logic |
|---|---|---|---|
| Infrastructural | Cloud contracts, “the stack,” and locked-in procurement | Normalizes surveillance and dependency as the price of participation | Territorial capture: the platform becomes the only space where learning can occur |
| Classification | Algorithmic prediction and linguistic detectors | Penalizes non-dominant registers; shifts the burden of proof to the student | Linguistic enclosure: defining “normality” to justify disciplinary regimes |
| Epistemic | Safety layers, response rules, and ranking logic | Encourages self-censorship; shifts authority from disciplines to code | Paternalism: silencing diverse ways of knowing under the guise of “safety” |
| Labor | Uncompensated “repair labor” and quality assurance | Diverts public capital, teacher time, from pedagogy to system stabilization | Extraction: public institutions provide the raw labor to de-risk private products |
Note. Table extracted from Moravec, J. w. (2026;), “Beyond the digital enclosure: AI, coloniality, and the pursuit of educational sovereignty”. Learning Futures and Emerging Technologies, 34(2). https://doi.org/10.1108/LFET-01-2026-0004
It is important to note that I do not argue against the use of AI in education. Rather, I argue that adoption without meaningful institutional control can weaken educational agency. A school system may appear “AI ready” while remaining unable to audit system behavior, contest classification thresholds, access usable data, review changes, recognize repair labor, or involve educators and students in decisions about continued use.
For this reason, the article frames educational sovereignty as a practical test of AI governance. Sovereignty means that educational communities can inspect, question, contest, and revise the systems that shape learning. This requires more than policies for users. It requires procurement rules, rights to audit, usable data access, appeal processes, visible system changes, recognition of repair labor, and meaningful participation by educators and students.
AI governance cannot stop at responsible use in educational environments. Classroom guidance may manage conduct, but it cannot govern platform power. If education is to remain a public good, institutions must govern not only how AI is used, but also how AI systems enter, classify, mediate, and reshape educational life.
