Michael Rowe

AI literacy

note5 min read

AI literacy is six capabilities held together by judgement

When people ask “are you AI literate?”, they often mean “can you write effective prompts?” That’s like asking if someone is information literate and only checking whether they can use Google. AI literacy takes technical knowledge, operational skill, ethical awareness, and critical thinking together, and what holds them together is contextual judgement.

What is AI literacy?

One-sentence definition: The multidimensional capability to recognise AI systems, understand how they work and where they fail, critically evaluate their outputs, understand the ethical implications of using them, and judge when and how AI serves genuine goals.

Using AI and creating with it belong to the six dimensions too, but only as far as understanding needs them. Beyond that, doing them well as part of everyday work is AI fluency, the practice that literacy makes possible.

A person can understand how language models work and still have poor judgement about when to use one. Someone who writes elegant prompts may accept whatever comes back, and someone with a deep grasp of the ethics may never have used the tools enough to engage with them productively. Strength in one dimension doesn’t carry over to the others, which is why all six have to develop together.

The six dimensions

The six dimensions come from well beyond the AI literacy literature. They appear across the information, media, digital, and data literacy traditions, where the content changes from one domain to the next and the common architecture of literacy underneath stays the same.

  1. Access and recognition — understanding what AI systems are and identifying when they are relevant
  2. Critical evaluation — assessing AI outputs for quality, accuracy, reliability, and bias; understanding complementary errors
  3. Functional application — practical competence with prompting, tool selection, and structured engagement
  4. Creation and communication — generating meaningful outputs through collaboration with AI while maintaining distinctive voice
  5. Ethical awareness — understanding social, professional, and civic implications of AI engagement
  6. Contextual judgement — developing professional taste about when and how AI engagement serves genuine goals, and when it undermines them

The sixth dimension integrates the others: technical competence without judgement leads to careless use, and ethical awareness without practical competence leaves a person unable to act on it. See the AI literacy development framework for how these dimensions develop through embedded practice across three stages: substitution, adaptation, and transformation.

A final-year physiotherapy student asks a chatbot for a home exercise programme for a patient going home after a knee replacement. The prompt is good, which is functional application. The student notices that one exercise isn’t appropriate this soon after surgery, which is critical evaluation, and has left the patient’s details out of the prompt, which is ethical awareness. Contextual judgement is the question underneath all three: whether this patient needs a printed sheet at all, or ten minutes of the student’s time going through the exercises with them. A student strong on the first three and weak on the fourth produces a safe, well-written handout for a patient who needed a conversation.

Developmental character

AI literacy is neither binary nor universal — it develops through increasing sophistication and differs across domains. Someone can be highly AI-literate in a research context while still developing literacy in clinical or supervision settings. Two things follow. Literacy can’t be assessed at a single point in time, and what counts as meaningful AI engagement varies with context and purpose. Developing AI literacy sets out how the progression usually unfolds and what makes it possible.

AI literacy and AI fluency

AI literacy is the foundation that AI fluency builds on. Literacy is mostly understanding: recognising AI systems, knowing how they work and where they fail, evaluating what they produce, and seeing the ethical questions they raise. Fluency is the practice of working with them, which the AI Fluency Framework (Dakan et al., 2025) describes as four competencies: delegation, description, discernment, and diligence. The six dimensions above already reach into that practice through functional application and creation, and contextual judgement is close to what the framework calls delegation, so the two overlap. The difference is in what each asks. A literacy framework asks what a person needs to know and be able to judge, and a fluency framework asks how well they work with AI on real tasks, which is why someone can be thoroughly literate and still be a hesitant or careless user.

In the knee replacement example, the student’s understanding of why a chatbot might suggest an exercise that’s unsafe so soon after surgery is literacy. Writing the prompt, catching the exercise, and deciding whether the patient needed a handout at all are fluency, and the student needs both.

Sources

Continue reading

  • The common architecture of literacy

    The six dimensions of literacy that information, media, digital, data, and AI literacy frameworks share, from access and evaluation to contextual judgement, and why that means AI literacy builds on what educators already teach.

  • AI literacy development framework

    A framework for embedding AI literacy development into existing modules and courses, enabling students to develop AI capability while learning disciplinary content.

  • Developing AI literacy

    Developing AI literacy takes more than regular use. How people move from substitution through adaptation to transformation, why many stop at the first stage, and the three conditions that make progression possible: foundational orientation, authentic use, and deliberate reflection.