ChatGPT for Teachers: A Practical Classroom Guide [2026]ChatGPT for Teachers: A Practical Classroom Guide: practical strategies and classroom examples for teachers

Updated on  

May 22, 2026

ChatGPT for Teachers: A Practical Classroom Guide [2026]

|

March 2, 2026

ChatGPT is a large language model that generates human-like text responses to natural language prompts.

ChatGPT is a large language model that generates human-like text responses to natural language prompts. For more on this topic, see 10 ai prompts every teacher. For teachers, it works like a teaching assistant that's always there. It can write lesson plans, create different resources, make assessment questions, simplify hard texts, and give quick feedback on student work. The key skill is learning to prompt it effectively.

Key Takeaways

  1. ChatGPT can reduce drafting time for some teacher tasks, but every output still needs professional review. DfE guidance frames teacher-facing generative AI as a possible support for resource creation, planning, feedback and administration, while warning that outputs can be inaccurate, biased, out of date or unsuitable. Treat AI as a drafting assistant, not as an evidence source or teaching decision-maker.
  2. Effective prompting is the cornerstone of generating high-quality, classroom-ready resources with ChatGPT. The quality of AI-generated educational materials is directly proportional to the specificity and pedagogical alignment of the input prompts, mirroring the principles of effective instructional design (Reigeluth, 1983). Teachers must leverage their expertise to craft prompts that ensure the AI produces content genuinely suitable for their learners' learning needs and curriculum objectives.
  3. ChatGPT can help draft differentiated materials and feedback templates, but it does not remove teacher judgement. Use it to create first drafts, alternative explanations and prompt variations, then check the result against the learner, the curriculum, accessibility needs and the evidence base for feedback and differentiation.
  4. Navigating the ethical implications and adhering to DfE AI guidance are paramount for responsible ChatGPT integration in schools. The responsible deployment of AI in educational settings necessitates careful consideration of data privacy, algorithmic bias, and the potential impact on learner agency and critical thinking, as extensively discussed in the field of digital education ethics (Selwyn, 2019). Adherence to national guidelines ensures that AI tools are used to augment, rather than undermine, sound pedagogical practice and learner safeguarding.

This guide focuses on practical applications, not theory. The examples are classroom-style prompts that teachers can adapt, test and review locally. The goal is to reduce low-value drafting time while keeping professional judgement, safeguarding and subject expertise at the centre of teaching.

ChatGPT: Your AI Teaching Assistant infographic for teachers
ChatGPT: Your AI Teaching Assistant

Essential ChatGPT Uses for Teachers

Lesson Planning and Resource Creation

For many teachers, the most practical use of ChatGPT is first-draft resource generation. It can speed up routine drafting, but the teacher still needs to check subject accuracy, curriculum fit, accessibility and safeguarding before anything reaches learners.

Starter activities: "Create a 5-question retrieval practice starter for Year 9 Biology on cell division. Include 2 questions from last lesson, 2 from last week, and 1 from last half term. Use a retrieval grid format with 1-mark recall, 2-mark explain, and 3-mark apply columns."

Differentiated worksheets: Make three versions of a fractions worksheet for Year 5 learners. Version A simplifies content with visuals and sentence starters. Version B matches standard curriculum targets. Version C extends learning with reasoning and problem-solving (Tomlinson, 2014).

Vocabulary resources should produce Tier 2 word lists. Focus Year 7 learners on Norman Conquest history. Resources should give word definitions. Provide historical example sentences. Identify common learner misunderstandings.

Create ten multiple choice electricity questions for GCSE Physics. Each question needs four answer options. Ensure each wrong answer shows a common learner misconception. Explain the misconception each incorrect option targets (Black & Wiliam, 2018). This helps identify gaps in learner knowledge (Christodoulou, 2017; Petty, 2014).

Marking and Feedback

ChatGPT can draft feedback on student work, though the teacher must always review and personalise the output.

Generating feedback templates: "I teach Year 10 English. A student has written a paragraph analysing the character of Lady Macbeth. They identified one technique (metaphor) but did not explore its effect on the reader. Write feedback that: acknowledges what they did well, identifies the gap, and provides a specific next step with a sentence starter."

Rubric-based assessment: "Here is a student's response to a GCSE Geography question about urbanisation. Mark it against the AQA mark scheme (4 marks: 1 for identification, 1 for description, 2 for explanation). Explain where they gained and lost marks."

Whole-class feedback scripts: "Based on these common errors from my Year 8 Maths class on simultaneous equations, write a whole-class feedback script I can use at the start of the next lesson. Focus on the misconception about negative coefficients."

Simplifying and Adapting Texts

For SEND students and those learning English as an additional language, ChatGPT can adapt complex texts while preserving meaning.

Motivation impacts learning. Deci and Ryan's (1985) self-determination theory links motivation to achievement. Learners need autonomy, competence and relatedness (Ryan & Deci, 2000). Intrinsic motivation improves learning outcomes, says Reeve (2012). Teachers should foster intrinsic motivation for better results (Pink, 2009).

Creating visual supports: "Convert this written explanation of the water cycle into a numbered step-by-step list with suggested simple icons for each step. This is for a student with autism who learns better from pictures than written text."

Sentence stems and frameworks: "Create scaffolding sentence starters for a Year 7 student writing a persuasive letter. The student has strong ideas but struggles to structure paragraphs. Provide a framework with connectives built in."

ChatGPT Prompting Strategies for Teachers

The CRAFT Framework

Teachers get better results when they structure prompts using CRAFT:

| Element | Meaning | Example |

|---------|---------|---------|

| C ontext | Your teaching situation | "I teach Year 8 mixed-ability History" |

| R ole | What you want ChatGPT to be | "Act as an experienced KS3 History teacher" |

| A sk | The specific task | "Create a knowledge organiser for the English Civil War" |

| F ormat | How you want the output | "Format as a single A4 page with key dates, people, and vocabulary" |

| T one | The style or audience | "Use language accessible to 12-13 year olds, UK English" |

A complete CRAFT prompt: "You are an experienced KS3 History teacher (R). I teach Year 8 mixed-ability classes in a UK secondary school (C). Create a knowledge organiser for the English Civil War covering causes, key events, and consequences (A). Format it as a structured table with three columns: Key People, Key Events, Key Vocabulary. Maximum 20 items total (F). Use clear, concise language that a 12-year-old would understand (T)."

Five Rules for Teacher Prompting

Rule 1: Specify the year group and subject. "Write questions for students" produces generic output. "Write questions for Year 6 students studying the Viking invasion of Britain" produces usable resources.

ChatGPT works well with pedagogy names. For example, request "Bloom's Taxonomy questions". Or ask for "dual coding resources", or "interleaving practice sets". This grounds the resources in research (Bloom, 1956; Paivio, 1971; Rohrer, 2012).

Rule 3: Specify what you do not want. "Do not include generic praise. Do not use American English. Do not suggest activities requiring technology we do not have."

Rule 4: Iterate, do not start over. If the first response is close but not right, refine: "Good, but make the questions harder. Add questions that require students to compare two events rather than just recall." Building on existing output is faster than rewriting prompts from scratch.

Rule 5: Always review the output. ChatGPT generates plausible but sometimes inaccurate content. Every fact must be checked. Every resource must be reviewed through a teacher's professional lens before it reaches students. This is not optional.

Subject-Specific ChatGPT Teaching Methods

English

Exemplar tasks ask learners to write paragraphs. For instance, "Analyse power in Macbeth Act 1 Scene 7". Ask for a PEE structure and specify the terminology you want learners to practise, such as soliloquy, metaphor and stage direction.

Write VIPERS questions for 'Holes', Chapter 3 by Sachar. Include two retrieval questions. Also, write two inference questions. Add one vocabulary question, and one prediction question. (Sachar, date unknown).

Grammar starters: "Create a 5-minute grammar starter on apostrophes for possession. Include 5 sentences where students must put apostrophes in the right place. Make sure to include examples with plural nouns."

Maths

Worked examples: "Create a worked example showing how to solve a quadratic equation using the formula. Include common errors students make at each step and how to avoid them."

Learners answer ten area questions. They first find standard triangle areas. Next, learners calculate height and use coordinate grids (Watson & Mason, 2006). Finally, they solve compound shapes, increasing difficulty (Bjork & Bjork, 1992).

Write five true or false questions on probability. Focus these questions on common Year 8 learner errors. For each question, explain the relevant misconception. This helps teachers diagnose misunderstandings. Address frequent problems learners face.

Science

Create a KS3 enzyme practical on temperature's effect. List equipment and a safe method. Predict results and add three assessment points.

Exam question generation: "Write a 6-mark GCSE Biology question on natural selection in the style of AQA. Include a mark scheme with acceptable answers for each mark."

Humanities and Languages

Provide learners with source analysis help for WW1 posters. Include questions about origin, said Wineburg (1991). Ask learners to analyse the poster's content. Prompt learners to assess reliability and purpose, as argued by Seixas (2004) and Lévesque (2008).

MFL sentence builders: "Create a sentence builder for Year 9 French on the topic of holidays. Include: time phrases, opinions, activities, and connectives. Ensure the grammar builds from simple present to past tense."

ChatGPT for Teachers: A Practical Classroom Guide [2026] infographic showing strategies for ChatGPT, prompting, and CRAFT framework for teachers
The Teacher's ChatGPT Toolkit

ChatGPT Limitations Teachers Must Know

Limitations Teachers Must Understand

It generates plausible nonsense. ChatGPT does not know facts. It predicts which words are likely to follow other words. This means it can produce confident, fluent text that is factually wrong. A teacher who uses ChatGPT to generate a History timeline without checking every date risks teaching inaccuracies.

It does not understand your students. ChatGPT cannot see that Priya needs more scaffolding on inference questions, or that Marcus is bored because the material is too easy. The teacher's relational knowledge of each student remains irreplaceable.

It cannot replace professional judgement. A prompt asking "Should I move this student to the higher group?" asks ChatGPT to make a decision it is not qualified to make. Use it for resource generation, not pedagogical decision-making.

ChatGPT poses data protection risks. Always protect learner data. Do not input learner names or assessment data. Never share personally identifiable details. Use anonymised examples such as "A learner wrote..."

Comparing ChatGPT with Educational AI Tools

AI tools change quickly, so this guide avoids fixed feature and pricing claims. Check OpenAI's current ChatGPT plan page for live ChatGPT features and pricing, and check your school-approved provider list before choosing any tool.

For organisation-wide use, managed education or business deployments generally need administrator controls, privacy settings and procurement review rather than individual accounts. OpenAI's ChatGPT Edu page explains its education deployment option; UK schools should still check DfE data-protection and safeguarding guidance before using any generative AI system with pupil data.

Ethical Use in Schools

School Policy Considerations

Schools need a clear AI policy covering:

Staff use: Which tasks can teachers use ChatGPT for? Most schools permit resource generation and planning support. Some restrict its use for writing reports or references.

Student use: Are students permitted to use ChatGPT? If so, for what? The most effective policies teach students to use AI as a tool while maintaining academic integrity. See also: Ai academic integrity.

Always protect learner data. Do not input learner names or assessment details. Anonymise any prompt with learner work. Staff should follow DfE data-protection guidance for generative AI in schools and local policy before putting any content into an AI tool.

Attribution: Should AI-generated resources be labelled? Many schools require a note: "This resource was created with AI assistance and reviewed by [teacher name]."

Teaching Students About AI

The best approach treats AI literacy as a curriculum skill. Students need to understand:

  • What AI can and cannot do (it generates text, it does not understand meaning)
  • How to evaluate AI output (check facts, identify bias, assess quality)
  • When AI is appropriate (research support, brainstorming) and when it is not (assessed work, unless specified)
  • How AI connects to critical thinking (AI makes the ability to evaluate and question more important, not less)

Structural Learning Thinking Framework helps assess AI output. Learners use coloured cards for specific tasks. Red "Evaluate" cards assess AI quality. Blue "Compare" cards check AI answers against textbooks. Orange "Target" cards find incorrect AI vocabulary.

ChatGPT for Teachers: A Practical Classroom Guide [2026] infographic showing the steps to Prompting, AI Assistant, and CRAFT Framework for teachers
5 Prompting Secrets for Better Results

Meeting DfE AI Guidance Requirements

DfE guidance asks schools and colleges to weigh benefits against risks, keep safety as the priority, protect personal data and use professional judgement when AI produces content. Use the current DfE generative AI policy paper and DfE data-protection guidance for generative AI in schools as starting points, then apply your local safeguarding, assessment and procurement rules.

Start with an agreed AI use case and a simple risk record: what the tool is being used for, whether any personal data is involved, how outputs will be checked, and who is accountable for the final resource. For open generative AI tools, avoid entering identifiable learner details.

Teachers should check AI outputs for bias, factual accuracy, curriculum fit and age suitability. AI can produce fluent but unreliable text, so final teaching decisions remain with the professional and the school.

How to Set Up ChatGPT for Teaching

Getting started with ChatGPT takes less than five minutes, but choosing the right route matters for classroom use. Visit chatgpt.com or your school-approved route. Account types, models, features and prices change, so check OpenAI's current plan page rather than relying on a fixed model or monthly-price claim. Use a school-managed account where your organisation requires one.

Protect learners: don't enter names or personal data in ChatGPT. Use "Learner A" instead. Check your school's AI policy with IT staff. If your school has a managed or closed AI tool, follow the agreed data-protection route before using any pupil information.

Once logged in, the interface is straightforward: type your prompt in the text box and press enter. Start with simple requests like 'Create five multiple choice questions on photosynthesis for Year 7' before attempting complex tasks. Save successful prompts in a document for quick reuse; many teachers build prompt libraries organised by subject and task type. For example, a Year 3 teacher might save 'Generate a word problem involving addition and subtraction within 100, using everyday classroom objects' as a template.

Use ChatGPT via a bookmark for quick lesson planning. The mobile app is good for generating ideas on the go. Desktop is easier for resource formatting. ChatGPT saves your history for later use.

Real-World ChatGPT Use Cases for Teachers

UK teachers use ChatGPT in new ways daily, going beyond resources. It can help with low-risk drafting for behaviour notes and parental communication, but teachers should separate drafting support from sensitive decisions that need human judgement.

Teachers can use ChatGPT to draft behaviour-report language or email structures from anonymised prompts. Keep incident details generic, then add verified facts and professional judgement outside the AI tool.

Teachers use ChatGPT for cross-curricular planning, saving time. They prompt: “Connect Romans with maths, English, and art for Year 3; include objectives." The AI generates a framework, which teachers refine to suit their learners' needs.

ChatGPT can help translate school terminology into plain-English parent messages. Teachers might ask it to explain "fronted adverbials" simply and suggest three practice activities, then personalise the message and remove anything that does not fit the family context.

ChatGPT effective methods for Effective Teaching

Use specific prompts to get useful ChatGPT outputs. Teachers in UK schools use these prompt structures daily. This makes ChatGPT a ready-to-use tool. Good prompts give better resources for learners.

Give clear context, like "Year 7 photosynthesis worksheet, 25 mixed ability learners, 3 with dyslexia". Include light reactions, diagrams, and short questions. Specific requests usually produce more usable drafts, and differentiation works best when the teacher specifies the learning aim, learner needs, supports and extension routes rather than asking for generic "easy, medium and hard" versions.

Divide complex tasks. Ask ChatGPT for learning objectives from your curriculum. Next, create questions for each objective. Lastly, request mark schemes. This staged method improves results and lets you refine each part.

Always review and adapt ChatGPT outputs before classroom use. While the tool can generate frameworks and initial drafts, your professional judgement ensures content matches your students' needs. Keep successful prompts in a document; a prompt library can save drafting time over repeated tasks, but treat saved prompts as starting points that still need subject, SEND and curriculum checks.

Remember that ChatGPT works best as a starting point, not an endpoint. Use it to overcome blank page syndrome, generate multiple options quickly, or create differentiated versions of existing resources. The tool amplifies your teaching expertise; it doesn't replace it.

Ready-to-Use ChatGPT Prompt Library

Here are ten ready-to-use prompts. Copy, adapt the year group and subject, and use.

  • Weekly starter pack: "Create 5 retrieval practice starters for Year [X] [Subject] on [Topic]. Each starter should have 5 questions: 2 from the last lesson, 2 from last week, and 1 from last term."
  • Knowledge organiser: "Create a knowledge organiser for Year [X] [Subject] on [Topic]. Include: key vocabulary with definitions, key facts, key people/dates, and a diagram or model."
  • Marking feedback bank: "Write 10 feedback comments for Year [X] [Subject] essays on [Topic]. 5 for students who need to improve (specific targets), 5 for students who need extension (next steps)."
  • Parents evening prompts: "I am preparing for Year [X] parents' evening. Write 5 sentence templates I can adapt for: a high-achieving student, a student making expected progress, a student who needs to improve behaviour for learning, a student with anxiety about tests, and a student with SEND."
  • Vocabulary pre-teaching: "Identify the 8 most important Tier 2 and Tier 3 vocabulary words in this text. For each, provide: definition, example sentence, and a common student misunderstanding."
  • Lesson closer: "Create 3 exit ticket questions for Year [X] [Subject] on [Topic]. One recall, one application, one that checks for the most common misconception."
  • Homework with choice: "Design a homework menu for Year [X] [Subject] on [Topic]. Include 6 tasks across different learning styles: one written, one visual, one verbal, one creative, one research, one practical."
  • Graphic organiser content: "Fill in a Venn diagram comparing [Concept A] and [Concept B] for Year [X]. Include 4 unique features for each and 3 shared features."
  • CPD summary: "Summarise the key findings from [research paper/book] in 200 words suitable for a 5-minute staff briefing. Include: one key takeaway, one practical classroom application, and one discussion question."
  • SoW overview: "Create a 6-week scheme of work outline for Year [X] [Subject] on [Topic]. For each week, suggest: key learning objectives, key vocabulary, suggested activities, and assessment opportunity."
  • Next lesson, take one low-risk planning task. Write a CRAFT prompt, review the draft, then spend your saved attention on the professional decisions ChatGPT cannot make: sequencing, misconception checks, learner needs and classroom relationships.

    Frequently Asked Questions

    Is ChatGPT allowed in UK schools and what are the data protection considerations?

    Some schools allow ChatGPT for planning, but policies differ. Check your school's AI rules before using it. Never enter learners' personal data into an open AI tool; use generic examples and anonymised samples, and follow your school's data-protection route.

    Can ChatGPT create resources that meet UK National Curriculum requirements?

    ChatGPT can draft resources linked to curriculum aims when you specify the year group, outcome and assessment purpose. It does not guarantee curriculum accuracy, so teachers should check outputs against the National Curriculum, exam-board specifications and school schemes of work before using them.

    How accurate is ChatGPT for creating subject-specific content and assessments?

    ChatGPT can produce plausible but incorrect activities, facts or mark schemes. Verify scientific data, dates, quotations, maths and assessment criteria before using outputs with learners. For important tests, check answers against exam-board rules.

    What are the best ways to prompt ChatGPT for differentiated learning materials?

    Differentiation prompts should clarify the learning aim, learner needs, and task format. Provide specifics: "Make three versions." Include visual supports for EAL learners, a standard version, and extensions for high-attaining learners. State year groups, reading levels, or SEND needs to pitch materials right.

    Can ChatGPT help with parent communication and school admin tasks?

    ChatGPT can draft emails and simplify jargon for parents, but a teacher should personalise the content, check the tone and avoid entering identifiable pupil information. Generic responses can seem cold to a learner's family.

    Further Reading: Verified Sources

    These sources support the article's claims about generative AI policy, data protection, differentiated drafting and feedback. Use each source for the specific point named below.

    Department for Education (2025). Generative artificial intelligence (AI) in education. View DfE policy paper

    Use this for teacher-facing generative AI opportunities and risks, including planning, resource creation, feedback, administration, emerging evidence and the need for professional judgement.

    Department for Education (2026). Data protection in schools: generative AI and data protection. View DfE guidance

    Use this for data-protection wording, especially anonymising prompts, checking with a DPO or IT lead, and understanding open versus closed generative AI tools.

    OpenAI. ChatGPT Plans. View current plan page

    Use this only for current ChatGPT features, plan names and pricing. Because these details change, avoid fixed model or price claims unless this page has just been checked.

    Chan, T. Y. H. and Wong, A. (2023). Effectiveness of Implementing Differentiated Instruction in the English for Specific Purposes Classroom in Hong Kong. View DOI record

    Use this cautiously for differentiated instruction in one Hong Kong tertiary ESP context. It does not prove a fixed engagement uplift in UK schools.

    Lui, A. M. and Andrade, H. L. (2022). The Next Black Box of Formative Assessment: A Model of the Internal Mechanisms of Feedback Processing. View DOI record

    Use this for the point that feedback is processed by learners, not merely delivered. AI-drafted feedback still needs teacher checks and learner-facing clarity.

    Hattie, J. and Timperley, H. (2007). The Power of Feedback. View DOI record

    Use this for feed up, feed back and feed forward questions. Do not use it as blanket proof that any AI-generated feedback improves attainment.

    Cognitive Science Platform

    Make Thinking Visible

    Open a free account and help organise learners' thinking with evidence-based graphic organisers. Reduce cognitive load and guide schema building dynamically.

    Create Free Account No credit card required
    Paul Main, Founder of Structural Learning
    About the Author
    Paul Main
    Founder & Metacognition Researcher

    Paul Main is an educator and metacognition researcher who founded Structural Learning in 2002. With a psychology degree from the University of Sunderland and 22+ years helping schools embed thinking skills, he bridges the gap between educational research and classroom practice. Fellow of the RSA and Chartered College of Teaching, with 128+ Google Scholar citations.

    More →

    Classroom Practice

    Back to Blog