Here is an interesting paper by Chiu and Rospigliosi (2026) in which they put six of the major frameworks side by side in a recent editorial in Interactive Learning Environments. This editorial extends Chiu’s (2025) earlier work separating AI literacy from AI competency, and pushes the conversation toward something the field has been avoiding: how teachers actually develop these competencies in real classrooms.
The Six AI Literacy Frameworks
The authors compare AI literacy and competency frameworks from UNESCO, the OECD/European Commission, Australia, China, the United Kingdom, and the United States. Each one targets a different audience. UNESCO, OECD, and China focus on students. The US Department of Labor and UK frameworks target workforce development. Australia’s framework addresses the entire school community, including parents and policymakers.
That audience choice shapes everything else: the tone, the structure, the level of prescriptiveness. China’s framework reads like a national curriculum directive, with grade-specific objectives across primary, junior secondary, and senior secondary stages within a “four-in-one” model of knowledge, skills, thinking, and values. The US framework reads like a workforce toolkit. UNESCO reads like an aspirational global reference. All six are about AI literacy, but they’re built very differently.
What They All Agree On
Despite the differences, four core commitments run through every framework. Ethics is universally non-negotiable, with bias detection, fairness, transparency, and responsible use treated as standard components. Human judgment is placed above AI, with the insistence that AI augment people. Literacy is defined as far broader than technical skill, including critical thinking, societal awareness, and the ability to evaluate AI outputs. Each framework also recognizes staged development across age, role, or job level.
Chiu and Rospigliosi point out that “All frameworks move beyond technical skills, defining literacy as encompassing critical thinking, societal awareness, values, and the ability to evaluate AI outputs critically” (p. 1004). That’s the consensus picture, and it’s worth naming clearly because too many discussions still reduce AI literacy to “knowing how to prompt.”
This connects to Hillman, Holmes, and Duarte’s (2025) rapid review of AI literacy frameworks for the Royal Society, which reached a similar finding: most frameworks have moved past the narrow technical-skills definition that dominated the early 2020s.
Where They Split
The frameworks split most sharply in their structural character. The authors describe UNESCO as aspirational and formative, China as prescriptive and strategic, OECD as integrative and assessable, the US framework as instrumental and vocational, the UK framework as operational and diagnostic, and Australia’s framework as protective and principled.
Those character differences matter for adoption. A teacher in Sydney working from Australia’s framework is doing safeguarding work. A teacher in Shanghai working from China’s framework is doing curriculum delivery. The same words, “AI literacy,” carry very different operational weight. Anyone trying to compare implementation across countries will need to start with that recognition.
The HCAP Bridge
The most consequential move in the editorial is the introduction of the Human-Centric AI Pedagogy (HCAP) framework. The argument is that the six frameworks describe what AI literacy should look like for different populations, but they don’t tell teachers how to actually develop those competencies. HCAP fills that gap from a teacher capacity perspective.
HCAP organizes teacher pedagogical knowledge into five domains: I-TK (technical mastery), I-CK (critical evaluation), I-PK (designing AI-enhanced learning), HAIC-K (structuring human-AI collaboration), and Ethics-K (responsible and equitable use). Each domain maps to specific components in the six frameworks.
The authors argue that “The HCAP framework provides the pedagogical ‘how’ that operates the six frameworks’ ‘what'” (p. 1005). That phrasing earns its place. Most policy frameworks I’ve covered, including UNESCO’s 2024 AI competency framework for students, define destination competencies without specifying the teaching journey. HCAP, which builds on the TPACK model that Mishra, Warr, and Islam (2023) extended for the ChatGPT era, is positioned as that journey map.
Where I’d Push
I agree with the central move. The six frameworks really do describe destinations clearly. The teaching journey is left vague. That gap, between “what students should know” and “what teachers need to do,” has needed naming for a while, and HCAP at least gives the conversation a vocabulary.
Where I’d push is on whether HCAP, as currently described, is concrete enough to actually function as that bridge. The five domains read like a TPACK-style conceptual scaffold, not a usable instrument for teacher development. A classroom teacher reading I-TK, I-CK, I-PK, HAIC-K, and Ethics-K needs substance, not acronyms. They need worked examples, classroom routines, and assessment design that match each domain. The editorial does not provide these. Whether the standalone HCAP paper does is a separate question.
The other thing worth saying: a global comparison is useful, but six frameworks are also six policy environments, six funding structures, and six teacher development systems. Cross-national synthesis is a starting point but not a roadmap.
The structured comparison is the value here. So is the move from policy frameworks to teacher pedagogy. The implementation work is what comes next. Frameworks define destinations. Teachers walk the journey. We need both, and we need them connected.
References
- Chiu, T. K., & Rospigliosi, P. A. (2026). Six global frameworks for human-centred AI literacy and competency: comparative analysis and a way forward. Interactive Learning Environments, 34(3), 1003-1005. https://doi.org/10.1080/10494820.2026.2648342
- Chiu, T. K. F. (2025). AI literacy and competency: Definitions, frameworks, development and future research directions. Interactive Learning Environments, 33(5), 3225-3229. https://medkharbach.com/ai-literacy-frameworks-2/
- Hillman, V., Holmes, W., & Duarte, T. (2025). A rapid review of AI literacy frameworks, with policy recommendations. A report prepared for the Royal Society. London: The Royal Society.
- Mishra, P., Warr, M., & Islam, R. (2023). TPACK in the age of ChatGPT and generative AI. Journal of Digital Learning in Teacher Education_, 39(4), 235–251. https://doi.org/10.1080/21532974.2023.2247480
- UNESCO. (2024). AI competency framework for students. United Nations Educational, Scientific and Cultural Organization. https://doi.org/10.54675/JKJB9835