I advocate for AI literacy in classrooms without apology. My concern is that what we’ve been teaching as AI literacy might already be inadequate for the moment we’re in. A new framework from Hauck, Moore, and Wright (2025) at The Open University makes that case in a way I find hard to argue with.
Their argument is that AI literacy as we commonly teach it misses something important: the question of power. Those questions include who benefits from these systems, who gets marginalized, and how educators decide which outputs to trust. Their framework for Critical AI Literacy answers those questions with an explicit equity lens, and I think it deserves close attention.
What Critical AI Literacy Actually Means
Hauck and colleagues build their framework on top of Long and Magerko’s (2020) well-known definition of AI literacy, but they argue it’s incomplete. Standard AI literacy teaches people to evaluate AI technologies, collaborate with them, and use them as tools. All of that is useful as far as it goes. It’s also insufficient if the people shaping and reviewing these tools ignore the question of whose knowledge counts and whose gets pushed aside.
Critical AI Literacy adds that missing dimension. The authors draw on Darvin’s (2017) work in Critical Digital Literacy to frame literacy itself as a social practice, something individuals do in context, not a fixed skill set they possess. This shift changes how the skill actually gets taught. A competency checklist treats AI literacy as learnable once and then checked off. A social practice treats it as ongoing, context-bound, and always political.
The centerpiece is what the authors call the EDIA lens: equality, diversity, inclusion, and accessibility. Every skill, every prompt, every output evaluation in the framework runs through this lens. When a student generates an image with AI, the question isn’t just “does it look good?” It becomes “who is represented, who is missing, and what assumptions is the model making?”
Hauck et al. take this logic further by connecting it to what they call epistemic injustice. LLMs, by their training and scale, tend to amplify dominant voices and flatten minority ones. Students who can’t recognize that pattern are not just missing a technical skill. They’re being trained to accept a narrower view of knowledge as the neutral default.
The Six Components of the Framework
The framework organizes Critical AI Literacy into six components arranged around the EDIA lens: AI concepts and applications, Learning and Teaching with AI, AI creativity, AI ethics, AI in society, and AI careers.
The first two are what most AI literacy frameworks cover: knowing what AI is, how it works, and how to use it in learning contexts. I’ve walked through similar territory in my reading of the UNESCO (2024) AI competency framework for students. What’s interesting here is how EDIA gets woven into each of those basic skills. Students aren’t just learning to identify a hallucination. They’re learning to ask whose knowledge gets treated as reliable and whose gets flagged as suspect.
The other four components are where the framework separates from typical AI literacy checklists. AI creativity asks students to run the same prompt multiple times across different tools to detect cultural bias, then engage in dialogue with the model to challenge those biases. That’s a specific, teachable move. It also treats creativity as iterative, which aligns with what Kalantzis and Cope (2025) have argued about literacy in the age of AI as being a design practice.
AI ethics, AI in society, and AI careers push the framework past individual classroom skills into systemic thinking. Students are asked to consider the carbon footprint of AI use, the labor exploitation behind model training in the Global South, the digital divide that keeps AI access uneven across populations, and the concentration of AI power in a small number of companies. These are uncomfortable questions that most classroom AI frameworks skip, and I think the authors are right to put them on the table.
Where This Framework Does Something Different
Most AI literacy frameworks I’ve covered stay inside a specific pedagogical comfort zone. They teach students how to use AI responsibly, how to evaluate outputs, how to cite AI assistance properly. All of that is useful and necessary. But they rarely step outside the question of individual competence to ask structural questions about AI as a social phenomenon.
Hauck and colleagues don’t do that. They put power at the center, and they mean it. Their advanced version of Critical AI Literacy connects directly to social justice-oriented action and change, drawing on Jiang and Gu (2022) and Mirra and Garcia (2020). The advanced framework asks how AI might help redress power imbalances, not just reproduce them. That’s a rare move in AI education writing.
It echoes arguments I’ve engaged with in Roe, Furze, and Perkins’s (2025) digital plastic metaphor, where the authors treat AI-generated text as a social substance with environmental and ethical weight, far removed from neutral information. Both papers refuse to treat AI as a politically empty tool.
Where the Framework Could Go Further
The framework is strong conceptually, but it’s a version 0.1 document, and that shows in its practical examples. Most of the guidance stays at the level of “educators should encourage students to consider X” or “students should engage with Y.” The specific activities, assignments, and rubrics that would turn these competencies into actual classroom practice aren’t there yet.
The authors acknowledge this and invite contributions to build a bank of concrete examples, which is fair, but a teacher reading the framework today still has to do most of the translation work themselves.
I also think the framework could do more with institutional context. It’s explicitly grounded in The Open University’s policies and strategic plans, which makes sense for an OU document, but raises questions for anyone trying to apply it elsewhere. A K-12 teacher in a small district won’t have the same institutional scaffolding, the same EDIA vocabulary, or the same policy alignment. The framework could travel better with some indication of how to adapt it when the institutional context is less supportive.
These are friendly criticisms of a strong piece of work. The framework’s core argument, that AI literacy has to include power analysis or it’s not really critical, is one I agree with fully.
What Teachers Can Try Right Now
Something I’ve been asking teachers to do in workshops is run the same prompt across two different chatbots and compare the outputs with their students. Whose faces appear in the generated images? Does the model default to certain names? When it writes dialogue, what dialects come through? This is a five-minute activity that puts the EDIA lens into action without requiring anyone to read a theoretical paper.
Another small move: after students use AI for any assignment, ask them to write a two-paragraph reflection on what the tool helped with, what it missed, and whose perspective was underrepresented in the output. This forces metacognition about AI’s limits in a way that tool-use alone never does. Both of these are small activities, but they start the muscle-building that Critical AI Literacy requires. I have a teacher-facing guide with multiple critical thinking activities to use in AI-enabled assignments.
I used Claude to create this one-page poster capturing the main ideas behind EDIA’s framework. You can download it here.
Final Thoughts
Critical AI Literacy, as Hauck and colleagues define it, raises the bar. It asks teachers and students to treat every AI interaction as a chance to ask who wins, who loses, and what gets taken for granted. That’s more work, but it’s the right work. The version of AI literacy that just teaches prompt-writing and hallucination-spotting is already behind the curve. The field needs to move toward Critical AI Literacy next.
References
- Darvin, R. (2017). Language, ideology, and critical digital literacy. In. S. Thorne, & S.
May (Eds.) Language Education and Technology, Encyclopaedia of Language
and Education (vol. 9, pp. 17–30). Springer. - Hauck, M., Moore, E., & Wright, C. (2025). A framework for the learning and teaching of critical AI literacy skills (Version 0.1). The Open University. https://www.open.ac.uk/blogs/learning-design/wp-content/uploads/2025/01/OU-Critical-AI-Literacy-framework-2025-external-sharing.pdf
- Jiang, L., & Gu, M.M. (2022). Toward a Professional Development Model for Critical
Digital Literacies in TESOL. TESOL Quarterly, 56(3), 1029–1040. https://doiorg.libezproxy.open.ac.uk/10.1002/tesq.3138 - Kalantzis, M., & Cope, B. (2025). Literacy in the time of artificial intelligence. Reading Research Quarterly, 60, e591. https://doi.org/10.1002/rrq.591
- Long, D., & Magerko, B. (2020). What is AI literacy? Competencies and design
considerations. Proceedings of the 2020 CHI conference on human factors in
computing Systems, 1–16. https://doi.org/10.1145/3313831.3376727 - Mirra, N., & Garcia, A. (2020). “I Hesitate but I Do Have Hope”: Youth speculative
civic literacies for troubled times. Harvard Educational Review, 90(2), 295–321.
https://doi-org.libezproxy.open.ac.uk/10.17763/1943-5045-90.2.295 - Roe, J., Furze, L., & Perkins, M. (2025). Digital plastic: A metaphorical framework for Critical AI Literacy in the multiliteracies era. Pedagogies: An International Journal. 1(15). https://doi.org/10.1080/1554480X.2025.2557491
- UNESCO. (2024). AI competency framework for students. United Nations Educational, Scientific and Cultural Organization. https://doi.org/10.54675/JKJB9835