AI Literacy for Teachers: Beyond Prompting
AI literacy for teachers combines practical skill with the judgement to decide whether and how to use AI in school. Prompting can improve an instruction, but it does not show teachers what information is safe to share, whether an output can be trusted or when to ask for school support.
In this article7 sections
Five questions teachers can use in everyday work
Use these questions when planning staff learning or reviewing an everyday teaching task. They are prompts for professional judgement, not a scorecard.
- 01
Can I use this tool for this task?
Teacher, within the school's approved usesCheck that the tool and task are allowed, understand its limits and stop when the fit is unclear.
- 02
Can I trust this output?
TeacherCheck facts, context, bias, age appropriateness and curriculum fit before using the output.
- 03
Does it support the learning?
Teacher, with academic guidanceStart with the learning purpose and decide whether AI use, or non-use, supports it.
- 04
Can I share this information?
Teacher, with data protection or IT supportRecognise personal or sensitive information, follow school rules and ask before sharing uncertain material.
- 05
Who should I ask when I am unsure?
Relevant school leadRoute questions about policy, assessment, safeguarding, data, IT or novel uses to the right person.
Why prompting is not enough
AI literacy for teachers starts where prompt practice stops. It prepares teachers to decide whether a tool belongs in a task, whether the output is dependable enough to use, what information can be shared and when they need support from the school.
Prompting still matters. A clearer instruction can produce a more relevant first response. But consider what happens when a teacher wants to prepare a quiz, adapt an explanation or draft a parent message. The prompt does not tell them whether the tool is approved, whether the material fits the curriculum, whether personal information can be entered or whether AI supports the learning purpose.
Those are professional decisions, not prompt-writing techniques. UNESCO’s AI competency framework for teachers connects practical AI knowledge with ethics, pedagogy and professional learning. The European DigCompEdu framework also places digital competence within teaching, assessment, learner needs and professional practice.
Use the five questions in everyday work
The five questions become useful when teachers apply several of them to one ordinary task. The examples below are not fixed permissions. Each school still needs to define which tools and uses are approved.
| Everyday task | What the teacher checks | When school support is needed |
|---|---|---|
| Prepare a short quiz | Facts, curriculum fit and age-appropriate language | The tool or intended use has not been approved |
| Adapt an explanation | Whether the output supports the learning or removes useful thinking | Curriculum or use expectations are unclear |
| Draft a parent message | Accuracy, tone and context before anything is sent | Personal or sensitive information may be involved |
The same tool can be suitable for one task and unsuitable for another. A fluent response may also miss the school’s curriculum, terminology or expectations. This is why school-aware AI needs more than a prompt.
Check the output and the learning purpose
A polished response can contain a wrong date, a weak explanation, unsuitable language or an assumption that does not fit the class. Teachers should check important claims against reliable sources, apply subject knowledge and look for missing context or bias. If the material cannot be checked to the standard the task requires, it should not be used.
In England, Department for Education guidance says generated content can be inaccurate, biased, out of date, unreliable or taken out of context. It keeps responsibility for the quality and appropriateness of work with education professionals and their organisations.
The output is only one part of the decision. Teachers should begin with what pupils need to understand or practise. AI may help a teacher explore alternatives in one activity, but remove the explanation, recall, drafting or discussion another activity is meant to develop. Choosing not to use AI should remain a credible professional decision.
Know when to stop and ask
If the tool, information or rule is unclear, stop before submitting anything and ask the person who owns the decision.
Removing a pupil’s name may not remove every identifying detail. Teachers should check both the approved tool and the approved use before entering information. UNESCO’s guidance for generative AI in education and research and the European Commission’s ethical guidelines for educators place privacy and human oversight at the centre of responsible use. In England, the DfE’s data-protection guidance tells schools to use approved tools, protect personal data and involve the data protection officer or IT lead when needed.
Teachers should not have to interpret every new use alone. Staff learning should name the people who can answer questions about curriculum, assessment, safeguarding, data protection, IT, security and policy. It should also show what teachers can decide within existing rules and which cases need a school decision.
For assessed work in England, Ofqual’s AI in coursework resources support a whole-school approach and consistent messages. Schools elsewhere should follow their own assessment rules.
What teachers decide and what schools own
Teachers need room to exercise professional judgement, but they should not have to carry every school-wide responsibility themselves.
Teachers decide within clear school rules
- Whether AI supports the educational purpose of the task.
- Whether the output is accurate, appropriate and suitable for the curriculum.
- Whether to reject an output, choose not to use AI or ask for clarification.
- When to stop because approval, information, assessment or safeguarding is unclear.
Schools set and maintain the boundaries
- Approve tools and define the uses that are allowed.
- Own policy, legal compliance, procurement, security and access.
- Protect privacy and safeguarding through clear processes.
- Name the people who can answer questions and resolve new cases.
The guide to leadership responsibilities before scaling AI takes this school-level responsibility further.
Review one staff-learning plan
Take one current staff-learning plan and compare it with the five questions. Look for one real gap, not a total score.
- Find the missing decision. The plan may teach prompting without explaining how to check an output, or explain what not to enter without showing whom to contact.
- Name the owner. Make clear what teachers can decide and who owns questions about curriculum, assessment, data, safeguarding, IT or policy.
- Change the next session. Add one approved-use instruction, checking routine, teaching scenario or visible route for asking for help.
Prompt training still deserves a place in teacher development. It helps staff give a tool clearer instructions and improve the first response.
But a better response is not the same as a sound classroom decision. Teachers still need to know whether AI fits the task, whether they can trust the material, what information can be shared and when the school needs to decide.
A useful next step is to add one of those decisions to the next staff session. That moves professional learning beyond technique and towards confident judgement supported by clear school boundaries.
