AI glossary for kids and teens
This AI glossary explains 24 words about how AI works, in plain language for kids and teens aged 8 to 18 and the adults who help them. Each word has a one-line definition below and its own page with an everyday example, a deeper section for older readers and a short no-screen activity.
Last reviewed: 20 October 2026 6 min read
Find a term
24 terms shown
No term matches that. Try a shorter word, or pick a group.
A to Z
- AI agentAn AI system that takes actions to finish a task, using tools step by step, instead of only answering.↗
- AlgorithmA list of clear steps that solves a problem or gets a task done, and always comes to an end.↗
- Bias in AIWhen an AI system's results are unfair or lopsided, often working better for some groups than others.↗
- ChatbotA computer program you can talk to by typing or speaking, which reads your message and writes a reply.↗
- Computer visionThe field that teaches computers to make sense of pictures and video, starting from pixel numbers.↗
- Data privacyA say over information about you: who collects it, what they use it for and who they share it with.↗
- DatasetAn organized collection of examples, one row each, gathered for a purpose such as training or testing a model.↗
- Deep learningMachine learning with neural networks of many layers, where each layer builds on what the one before found.↗
- DeepfakeA fake picture, video or voice recording made with AI to look or sound like a real person.↗
- Generative AIAI that makes new text, pictures, sound or video by following patterns learned from many examples.↗
- HallucinationWhen an AI system states something false or made up as if it were true.↗
- LabelThe right answer attached to an example, such as "cat" on a photo of a cat.↗
- Large language modelAn AI model trained on a huge amount of text to predict the next token, so it can write.↗
- Machine learningA way of making a computer find its own rule from examples, instead of a person writing the rule.↗
- ModelWhat a computer builds from examples to turn an input into an output, such as a prediction.↗
- Natural language processingThe field of AI that helps computers work with human language, written or spoken.↗
- Neural networkA system built from small units in layers that weigh their inputs, add them up and pass results on.↗
- OverfittingWhen a model learns its training examples too closely and does badly on new ones.↗
- ParameterA number inside a model that training adjusts, such as a weight that sets how much an input counts.↗
- PromptThe message you give an AI system to tell it what you want: a question, instruction or text.↗
- Prompt injectionA trick that hides instructions in text an AI system reads, so it follows them instead of yours.↗
- Supervised learningTraining a model with examples that already have the right answer attached, then testing it on new ones.↗
- TokenA small piece of text, such as a word, part of a word or punctuation, that a language model uses.↗
- Training dataThe examples a model learns from, kept separate from the test data used to check it.↗
Start with these three
New to AI words? Read these in order. Each one builds on the one before.
- Algorithma list of clear steps that solves a problem. It is the starting point for everything else here.↗
- Machine learninginstead of a person writing the steps, the computer finds its own rule from examples.↗
- Neural networkthe kind of model behind much of modern AI, built from small units that weigh their inputs.↗
How machines learn
- AlgorithmA list of clear steps that solves a problem or gets a task done, and always comes to an end.↗
- Computer visionThe field that teaches computers to make sense of pictures and video, starting from pixel numbers.↗
- Deep learningMachine learning with neural networks of many layers, where each layer builds on what the one before found.↗
- Machine learningA way of making a computer find its own rule from examples, instead of a person writing the rule.↗
- ModelWhat a computer builds from examples to turn an input into an output, such as a prediction.↗
- Neural networkA system built from small units in layers that weigh their inputs, add them up and pass results on.↗
- ParameterA number inside a model that training adjusts, such as a weight that sets how much an input counts.↗
- Supervised learningTraining a model with examples that already have the right answer attached, then testing it on new ones.↗
Data and testing
- DatasetAn organized collection of examples, one row each, gathered for a purpose such as training or testing a model.↗
- LabelThe right answer attached to an example, such as "cat" on a photo of a cat.↗
- OverfittingWhen a model learns its training examples too closely and does badly on new ones.↗
- Training dataThe examples a model learns from, kept separate from the test data used to check it.↗
Language and generation
- AI agentAn AI system that takes actions to finish a task, using tools step by step, instead of only answering.↗
- ChatbotA computer program you can talk to by typing or speaking, which reads your message and writes a reply.↗
- Generative AIAI that makes new text, pictures, sound or video by following patterns learned from many examples.↗
- HallucinationWhen an AI system states something false or made up as if it were true.↗
- Large language modelAn AI model trained on a huge amount of text to predict the next token, so it can write.↗
- Natural language processingThe field of AI that helps computers work with human language, written or spoken.↗
- PromptThe message you give an AI system to tell it what you want: a question, instruction or text.↗
- TokenA small piece of text, such as a word, part of a word or punctuation, that a language model uses.↗
Safety and trust
- Bias in AIWhen an AI system's results are unfair or lopsided, often working better for some groups than others.↗
- Data privacyA say over information about you: who collects it, what they use it for and who they share it with.↗
- DeepfakeA fake picture, video or voice recording made with AI to look or sound like a real person.↗
- Prompt injectionA trick that hides instructions in text an AI system reads, so it follows them instead of yours.↗
How to use this glossary
- Start with the one-line definition. It is written so a 9-year-old can follow it, and the term's page opens with the same idea.
- Read the page together. Every term page has an everyday example, then "A little deeper" for ages 11 to 13 and "Going further" for ages 14 to 18.
- Do the "Try it". Each page ends with an activity of five minutes or less that needs paper, objects or people, not a screen.
- Follow the links. Related terms on each page lead to the next idea, so you can move from one word to the next in a sensible order.
- Check yourself. The Quick check on each page has two or three questions with answers.
Questions
It explains 24 words about how AI works, grouped into four ideas: how machines learn, data and testing, language and generation, and safety and trust. Each word has a one-line definition here and its own page with an example, deeper sections and an activity.
Start with algorithm, then machine learning, then neural network. An algorithm is a list of steps, machine learning is a computer finding its own rule from examples, and a neural network is the kind of model behind much of modern AI.
AI is the broad field of making computers do tasks that seem to need intelligence. Machine learning is one way to do that, in which the computer learns a rule from examples, and deep learning is machine learning with neural networks of many layers.
Both. Every term page opens with a definition a 9-year-old can follow, then adds a section for ages 11 to 13 and another for ages 14 to 18 that go further into the same idea.
No. Each term page has a "Try it" activity of five minutes or less that uses paper, everyday objects or other people. ---
Mechna Academy is a city of machines for curious children aged 8 to 18, opening first to ages 8 to 10. In each mission, a child guesses how a machine decides, then opens its Peek Hatch to check, changes one thing and tests it on a case it has never seen. Children learn how AI works this way, and they never chat with an AI.
For grown-ups: join the waitlistPlaces are limited. The waitlist is for parents and guardians 18 and older.