New arXiv paper redefines AI literacy: from consumer skills to epistemic power
Foucault and Freire walk into an AI classroom — literacy gets a rebellion.
A new paper from Brady D. Lund and Zoë Abbie Teel, posted on arXiv (2607.27547, cs.AI/cs.CY) on July 30, 2026, challenges the dominant approach to AI literacy. Most frameworks today emphasize technical competency and "responsible use" — but the authors argue this is just a dressed-up consumer orientation. It teaches people how to operate AI systems without questioning who built them, what assumptions are baked in, or who benefits from their spread. The result? Users gain proficiency but lose epistemic agency.
Drawing on Michel Foucault's power-knowledge theory and Paulo Freire's pedagogy of critical consciousness, the paper reframes AI literacy as a critical practice with three pillars: contextual use (understanding when and why AI is appropriate), critical interrogation (probing biases and structural assumptions), and participatory governance (having a seat at the table in shaping AI policy). The authors also connect unequal AI access to longstanding epistemic injustices, arguing that a real literacy framework must confront structural inequities. The takeaway: literacy isn't just about clicking the right prompts — it's about reclaiming power over knowledge production in an AI-driven society.
- Authors Lund & Teel argue current AI literacy education enforces consumer passivity, not critical thinking
- Proposes a 3-part framework: contextual use, critical interrogation, participatory governance
- Applies Foucault's power-knowledge and Freire's critical consciousness to AI education
Why It Matters
As AI shapes knowledge production, literacy must shift from usage skills to critical agency and governance — for everyone.