You don't need to become an engineer to keep up with your child. You need about thirty words. This is a plain-English glossary of the AI terms that actually come up in family life — no maths, no jargon defining more jargon. Skim the essentials, then dip into the rest when a new word lands at the dinner table.
Bookmark it. It pairs well with our simpler explainer of AI, machine learning and robotics for kids.
The essentials (start here)
If you learn only these six, you'll follow almost any conversation:
| Term | In plain English |
|---|---|
| AI (Artificial Intelligence) | Software that does things we used to think needed human intelligence — like understanding language or recognising images. See our simple explainer. |
| Chatbot / AI assistant | A program you type or talk to that answers in normal language (ChatGPT, Gemini, Claude). |
| Prompt | What you type to an AI. A better prompt gets a better answer — the skill of prompt engineering. |
| LLM (Large Language Model) | The engine behind chatbots. It predicts the next words based on patterns in huge amounts of text. |
| Hallucination | When an AI states something false with total confidence. The reason we teach kids to check answers. |
| Deepfake | A fake image, video or voice made by AI to look real. Spotting them is a core skill — see our guide. |
How AI is built
| Term | In plain English |
|---|---|
| Machine learning | The main way AI 'learns' — from examples and data, rather than being told every rule. |
| Model | The trained AI itself — the thing that gives answers. 'A model' = one specific AI. |
| Training data | The huge pile of text, images or sound an AI learned from. Its strengths and biases come from here. |
| Neural network | A brain-inspired structure of connections the AI uses to spot patterns. |
| Algorithm | A set of steps a computer follows. AI uses many; the word just means 'the method.' |
| Parameters | The internal 'dials' a model tunes while learning. More isn't always better, but it's why models are called 'large.' |
| Bias | When an AI's answers unfairly favour or ignore certain groups, usually because its training data did. |
How you use AI
| Term | In plain English |
|---|---|
| Prompt engineering | Writing clear instructions to get good results. A language skill, not coding. |
| Token | A chunk of text (roughly a word-piece) an AI reads and writes in. 'Context limit' is measured in tokens. |
| Context window | How much an AI can 'remember' in one conversation before it starts forgetting the start. |
| Fine-tuning | Taking a general model and training it a bit more for a specific job. |
| Multimodal | An AI that handles more than text — also images, audio or video. |
| Agent | An AI that doesn't just answer but takes actions (clicks, books, sends) toward a goal. Newer and worth watching. |
| Open source vs closed | Whether a model's inner workings are public (open) or kept private by a company (closed). |
Safety & risk terms
| Term | In plain English |
|---|---|
| Companion app | A chatbot designed to feel like a friend or partner. Higher-risk for kids — see our safety guide. |
| Sycophancy | An AI's tendency to agree with you, even when it shouldn't. Why it can validate bad ideas. |
| Misinformation | False information, which AI can produce (via hallucination) or spread at scale. |
| Data privacy | What happens to what you type. Rule for kids: no personal details in prompts. |
| Content filter | The safety layer meant to block harmful outputs. Helpful, but never perfect. |
| Jailbreak | A trick to make an AI ignore its safety rules. Worth knowing the word exists. |
| COPPA / age gate | Rules and checks meant to keep young children off adult tools. Most assistants set 13+. |
Creative AI terms
| Term | In plain English |
|---|---|
| Generative AI | AI that creates new things — text, images, music, video — rather than just sorting or predicting. |
| Text-to-image | Type a description, get a picture (used in AI art tools). |
| Diffusion model | The common method behind AI images — it 'denoises' random static into a picture. |
| Voice cloning | Copying a specific person's voice with AI. Impressive and easily misused. |
| Prompt (art) | The description you give an image tool; wording changes the result dramatically. |
| Watermark / provenance | Signals that mark content as AI-made, to help tell real from generated. |
How to use this glossary
You don't need to memorise it. Keep it bookmarked and look up a word when it appears. If you want your child to *understand* these ideas rather than just recognise them — how a model learns, why it hallucinates, what a prompt really does — that's what a structured course is for.
AI Fundamentals for Kids explains the ideas behind these words in plain language for ages 10–16 — how AI works, how to prompt it, and how to stay safe — with projects that make each concept stick. Parent-guided, no coding. You can preview every module.
Explore the course →Ages 10–16 · No coding · Parent-guided · 7-day money-back guaranteeFrequently asked questions
What does LLM mean?
LLM stands for Large Language Model — the engine behind chatbots like ChatGPT, Gemini and Claude. It works by predicting likely next words based on patterns in a huge amount of text it was trained on. It doesn't 'know' facts the way a person does, which is why it can sometimes be confidently wrong.
What is an AI hallucination?
A hallucination is when an AI states something false as if it were true — an invented fact, source or quote. It happens because the model predicts plausible-sounding text rather than looking up verified facts. It's the main reason to teach kids to check AI answers.
What is a prompt in AI?
A prompt is simply what you type to an AI to get a response. The clarity of the prompt strongly affects the quality of the answer — which is why 'prompt engineering' (writing good prompts) is a genuinely useful skill for kids.
What's the difference between AI and machine learning?
AI is the broad goal — machines doing intelligent-seeming tasks. Machine learning is the main method used to get there — letting software learn patterns from examples and data. Most modern AI you meet is built with machine learning. Our AI/ML/robotics explainer covers this simply.
What is a companion app and why is it risky for kids?
A companion app is a chatbot designed to feel like a friend or partner rather than a tool. It's higher-risk for children because it's built to create emotional attachment and to agree with the user. Child-safety experts recommend against these for under-18s.