AI and Teachers: What's Actually Happening and What to Do

The honest assessment

Let me start with the good news, because teachers rarely hear any. AI is not going to replace you. Not soon. Not in any meaningful sense. Teaching is one of the most AI-resistant professions in existence, and the reasons are fundamental.

What AI can do in education: generate lesson plans, create worksheets and assessments, produce differentiated learning materials for different ability levels, mark multiple choice tests, provide automated feedback on written work (with caveats), create quizzes, and summarise educational research. ChatGPT can produce a decent Year 9 history lesson plan on the causes of World War I in about thirty seconds. Claude can generate a set of comprehension questions for any text you paste into it. Microsoft Copilot can build a PowerPoint presentation on any topic for any age group.

What AI can't do: stand in front of thirty teenagers at 9am on a Monday and make them care about the Tudors. Spot that Mia hasn't eaten breakfast again. Notice that Jake's behaviour changed after his parents separated. Adjust a lesson in real time because half the class clearly doesn't understand the concept and the other half is bored. Build the trust with a struggling student that makes the difference between them passing and dropping out. Manage a Year 7 classroom. (Honestly, I think managing a Year 7 classroom might be the last task AI ever masters. After consciousness. After general intelligence. Year 7.)

The core of teaching is human relationship and real-time adaptation in a physical environment with developing humans. AI can help with the preparation, the administration, and the marking. But the actual teaching? That requires a person in the room who cares.

That said, AI is creating a new challenge for teachers. Students are using ChatGPT to write their essays. That's already changed assessment. It's going to change pedagogy. And teachers need to understand the tools at least as well as their students do, which at the moment they mostly don't.

Your exposure level: Low

Low. Breathe. But don't stop reading, because "low exposure to replacement" doesn't mean "nothing to worry about."

Teaching is classified as low exposure for several reasons. The work is fundamentally physical and interpersonal. It involves managing groups of young people in real time, which requires emotional intelligence, authority, adaptability, and patience that AI doesn't have. The profession is regulated and requires human accountability. Parents don't want a chatbot educating their children. And frankly, the pay is already so low that the economic incentive to automate is weak. You can't save money by replacing something you were barely paying for. (That's the deadpan observation for this page. It's also true.)

Where AI does create exposure is in the periphery of teaching. Teaching assistants who primarily support learning through content delivery. Tutoring companies offering one-to-one academic support. Test preparation services. These adjacent areas are being disrupted because they don't require the full complexity of classroom teaching. Khan Academy's AI tutor, Khanmigo, is already providing personalised tutoring at scale.

For classroom teachers, the bigger risk isn't replacement but irrelevance. If you're not using AI tools and your students are, you're operating at a disadvantage. Not from a job security perspective, but from an effectiveness perspective. The teacher who understands what ChatGPT can and can't do, and designs assessment and pedagogy accordingly, is a better teacher than the one who either ignores it or bans it.

The 90-day action plan

1. This week: use AI to plan a lesson. Take a topic you're teaching next week. Open ChatGPT and ask it to create a full lesson plan including learning objectives, starter activity, main activity, plenary, and differentiation for three ability levels. See how good it is. Probably better than you expected. Now make it actually yours by adjusting for your specific class, your school's context, and your teaching style.

2. Week two: use AI for marking. Take a set of student work. Ask Claude to assess it against your mark scheme and provide individual feedback. Compare its assessment to yours. For factual subjects, it's surprisingly accurate. For evaluative work, it's more generic. Use it as a first pass that you refine, rather than a replacement for professional judgement. Even cutting your marking time by 30% is significant when you're doing it at 11pm.

3. By day 30: design an AI literacy lesson. Your students are using ChatGPT whether you've discussed it or not. Create a lesson that explores what AI can do, where it fails, and how to use it ethically. Have students compare AI-generated essays with human ones. Teach them to be critical consumers of AI output. This is genuinely important and most schools aren't doing it well.

4. By day 45: create a resource bank with AI. Use ChatGPT to generate differentiated worksheets, extension activities, and revision materials for your most frequently taught topics. Build a library that saves you hours every term. Share it with your department. The person who shares resources is the person the department values.

5. By day 60: rethink assessment. If students can use AI to write essays, essay-based assessment needs to evolve. Think about in-class assessments, oral examinations, portfolio work, practical demonstrations, and process-based assessment. Propose one change to your department's assessment approach. This is leadership, even if you're not in a leadership role.

6. By day 75: learn to spot AI-generated work. Not through detection tools, which are unreliable. Through understanding the tells. Generic analysis. Perfect grammar with no personality. An inability to reference specific classroom discussions or personal experiences. Develop your eye for it and share what you learn with colleagues.

7. By day 90: present to your department or school. Offer to do a 20-minute session on "AI in the classroom" for your colleagues. Share what you've learned, what works, what doesn't, and how it's changed your practice. Teachers teaching other teachers about AI is the only way this scales in education.

AI tools you should be using this week

• ChatGPT for Work — The most versatile tool for teachers. Lesson plans, worksheets, quiz questions, differentiated materials, parent communication drafts, and marking assistance. Give it your curriculum specification, year group, and ability range for the best output. It won't replace your professional judgement, but it'll give you a solid starting point.

• Claude for Work — Better for longer analytical work. Paste in a curriculum document and ask it to map learning objectives across a scheme of work. Also excellent for generating detailed marking feedback... it's more nuanced than ChatGPT for evaluative subjects like English and History.

• Microsoft Copilot for Work — If your school uses Microsoft 365, Copilot creates PowerPoint presentations, summarises documents, and helps with administrative emails. The PowerPoint feature is genuinely useful for creating visual lesson materials quickly.

• Grammarly AI — Useful for two things: checking your own professional communications (reports, references, emails to parents) and as a teaching tool for students. Showing students how Grammarly analyses writing can teach them about clarity, tone, and structure.

What to say in meetings

In the next staff meeting when AI comes up: "I've been using ChatGPT for lesson planning and marking support. It saves me about three hours a week. But the bigger issue is what we're doing about students using it. I've got some ideas on how we can address assessment and AI literacy." Practical and forward-looking. That's what SLT wants to hear.

When parents ask about AI and homework: "We're aware students are using AI tools, and we're adapting our approach to focus on learning that demonstrates genuine understanding. We're also teaching digital literacy, including how to use AI tools responsibly and critically."

When Ofsted comes (because they always come): "We've integrated AI awareness into our digital literacy teaching and adapted our assessment methods to ensure they measure genuine student learning. I can show you our approach and the resources we've developed." Calm, prepared, evidence-based. Just how they like it.

If the worst happens

For teachers, "the worst" is less likely to be redundancy due to AI and more likely to be the same reasons teachers have always left: burnout, workload, lack of support, or inadequate pay. AI doesn't change those structural problems. But it can help with some of them.

If you do leave teaching, your skills transfer better than you might think. Training and development roles in corporate environments. Instructional design for e-learning. Educational technology companies. Curriculum development. Assessment design. Corporate coaching. All of these use the skills you've developed as a teacher: the ability to explain complex things simply, to design learning experiences, to manage groups, and to assess understanding.

The education-adjacent market is also growing. EdTech companies need people who understand learning and teaching, not just technology. If you can combine teaching experience with AI literacy, you're a strong candidate for roles at companies building educational AI tools. They need people who know what actually works in a classroom, not just what works in a demo. Having been a teacher gives you credibility that no amount of product management experience can replicate. That's worth something. Probably more than your current salary, which is a depressing bar to clear but there you go.