Start here · Foundations
AI Without the Jargon
You do not need to be technical to use AI well, and you do not need to love it either. You need enough understanding to make good decisions for your learners. This page explains what AI is in plain English, what it can and cannot do, and how to start safely.
What is AI?
Artificial intelligence describes computer systems that carry out tasks we associate with human intelligence — spotting patterns, recognising speech, translating, recommending and writing. AI is a broad field: many language models predict the next fragment of text to generate their output, but not every AI system works this way.
You already use AI every day without noticing: predictive text on your phone, spam filters in your email, route planning in maps, automatic subtitles and the recommendations on your streaming service.
The words you’ll meet
Machine learning — systems trained from examples rather than only fixed, hand-written rules.
Generative AI — produces new text, images, audio or video in response to your instructions.
Large language model (LLM) — a model trained on large amounts of language to generate or process text. Fluent output can still be wrong.
Prompt — the instruction you give the system.
The confident new assistant
Imagine a new assistant who has read an enormous amount but has never taught your class and never met your learners. They write quickly and sound completely sure of themselves; sometimes they make things up. You would check their work before it went anywhere near a learner. Treat AI output the same way.
| Often useful for | Do not rely on without rigorous checks |
|---|---|
| Drafting and rewording text | Facts, dates and references |
| Generating activity ideas | Learner-specific context |
| Adapting format and reading demand | Final assessment decisions |
| Organising and structuring materials | Emotional or safeguarding advice |
| Creating quiz questions you then verify | Anything where an error could cause harm |
The risks, honestly
- Hallucinations — AI can fabricate claims, statistics and sources with total confidence.
- Bias — outputs can reflect bias present in training data.
- Privacy — never enter personal or learner-identifying information into an unapproved tool.
- Copyright — uploading other people’s materials may breach copyright and licence terms.
- Safeguarding — learners may meet inappropriate content; know your reporting routes.
- Over-reliance and assessment integrity — AI must not produce evidence a learner is required to produce independently, and awarding-body rules apply. AI detectors alone do not prove misconduct.
Follow your organisation’s policy and approved tools at all times. This page supports professional judgement — it does not replace safeguarding procedures, employer policy or data protection advice.
Common myths
| The myth | The reality |
|---|---|
| “AI will replace teachers.” | AI cannot build relationships, read a room or keep a learner safe. Roles and tasks may change — but judgement, care and accountability stay with people. Nobody can promise teaching roles will never change. |
| “Confident means correct.” | Fluent, certain-sounding output is routinely wrong. Confidence is a writing style, not evidence. |
| “AI detectors prove cheating.” | Detectors are unreliable and produce false positives, especially for neurodivergent writers and speakers of English as an additional language. They are not proof. |
| “It’s too late to start.” | It isn’t. Starting small, with one task and an approved tool, is exactly how confident practice begins. |
| “Everyone must use AI.” | AI is one option, not a duty. Some tasks suit it; many don’t. Choice stays with you and your organisation. |
Your first week: a five-day plan
- Watch and read. See a demonstration of an approved tool and read your organisation’s AI policy.
- Draft something low-stakes. Use an approved tool to draft a generic class reminder — no personal data anywhere in it.
- Improve the prompt. Add audience, level and format, and compare the two outputs.
- Verify. Ask for five facts in a topic you know well, and check every one.
- Choose one small task for next week — and reflect on whether it helped.
Classroom examples
Real uses, checked by a human
Entry Level literacy: a teacher uses AI to simplify instructions, then checks and corrects them before use.
Level 2 ICT: a lecturer edits three spreadsheet formula starter ideas generated as a first draft.
Adult returners: a tutor translates an assessment brief into plain English, then checks learner understanding.
Leadership: a head of department checks a ten-bullet policy summary against the original document.
Troubleshooting
| Problem | Fix |
|---|---|
| Vague output | Add level, audience and time to your prompt. |
| American tone | Request British English explicitly. |
| Repetition | Set a length limit. |
| Unverifiable facts | Check against official sources before use. |
| Feeling overwhelmed | Start with one task. That is enough. |
Further reading and downloads
Start with the Research, Guidance & Evidence page for the official sources behind this Hub, then take a ready-made prompt from the Practical AI Toolkit.
Downloads for this page
AI Without the Jargon is Book 4 of the Back in the Room Education collection — currently in development and not yet published. Books 1–3 are available now. This free Hub is built from the same practical approach.
Page last reviewed: 7 November 2026 · Next review: by May 2027 · Guidance sources checked on this date — see Research, Guidance & Evidence.