Turning schools into AI testing grounds and sites of extraction: The strategic selectivities of IOs AI literacy frameworks

This paper critically examines how two major international organizations, the OECD and UNESCO, frame “AI literacy” in their key policy documents. Using text-mining and network text analysis, we compare their visions and the strategic selectivities they promote for reshaping national education systems. Our analysis, informed by Jessop’s strategic-relational approach, reveals a profound socio-technical shift: rather than integrating technology into society, these frameworks depict society being integrated into GenAI. In this new education landscape, students act as “cybernetic” co-producers of knowledge, continuously interacting with AI systems in feedback loops, while teachers are largely marginalized, reduced to roles of basic monitoring instead of pedagogical mediation. Similarly, educational institutions are emptied of their pedagogical role and reduced to legitimizing compliance. We argue these selectivities represent a radical departure from traditional educational technology policy. In these documents, GenAI is leveraged as a totalizing force that restructures classrooms around productivity and testing, while systematically obscuring crucial questions of teacher agency, institutional context, and critical engagement. Particularly in the OECD/EU framework, schools become both testing grounds and sites of extraction for GenAI systems, fundamentally altering educational priorities and actor roles in a world dominated by oligarchic powers.