What sits behind a tool that writes text
A route planner, a photo app and a chat assistant can all be sold as AI, but they do not work the same way. AI for Everyone is a short course for people without a technical background: what kind of software a text-writing AI tool is, why its answers can read well and still be wrong, and how to check them.
The course has three lessons. The first sorts software into three kinds: programs that follow rules people wrote, models that learned from data, and large language models, which are one kind of machine learning model. You label example apps and put AI, machine learning and LLMs in order from widest to narrowest. The second lesson shows how a language model writes: the text is cut into tokens, every possible next token gets a probability, and a temperature setting changes the pick. You read a probability table, estimate how many tokens a 1,200-word document is and explain why a probable sentence is not a checked one. The third is about hallucinations: picking out the checkable claims in an answer, finding a primary source for each and comparing the details.
Each lesson opens with a short explanation, then gives you six steps: you pick a label, put items in order, type a number or write a few sentences. When your answer is off, the AI tutor first asks a question about it, then gives a hint, then a partial step. Show solution is always available, and a short quiz before and after each lesson shows what changed.
This is one of the editorial courses in Ai Teacher. The problems and worked solutions were written in advance, and facts and definitions were checked against the public sources listed below. No real AI product is named in the examples.
What you'll learn
- Tell rule-based software, machine learning and large language models apart
- Explain how a language model picks the next token
- Check an AI answer against a primary source
Lessons
AI, machine learning and LLMs: what is what
Tokens and probabilities
Hallucinations and how to check an answer
A checking step from lesson 3

Tutor: In this made-up example, a chat assistant says Larkmoor Bridge is 410 metres long. The operator's technical report gives 401 metres. Is the claim confirmed, contradicted or not found yet?
You: Confirmed. It's close enough.

Tutor: Would a builder accept 410 metres when the plan says 401?
You: No. I found the source and it disagrees, so the claim is contradicted.

Tutor: Yes, contradicted. A detail that does not match the original record is wrong, even when it looks close.
Who this course is for
- You use chat assistants at work or at home and want to know what happens behind the answer.
- You have no technical background, and words like machine learning, LLM or token still feel vague.
- You have been caught out by a confident AI answer and want a simple way to check the next one.
- You would rather work through short tasks with feedback than read long explanations.
Questions about AI for Everyone
Do I need a technical background?
No. There is no code in the course, and the math is simple arithmetic with percentages. The steps ask you to label examples, put steps in order, work out a small sum or explain an idea in a few sentences.
Will this course teach me to write better prompts?
No. The three lessons explain what kind of software an AI tool is, how a language model produces text and how to check what it says. They do not cover prompts or any particular AI product.
Is the course free?
One lesson in Ai Teacher is free. After that, a subscription opens all lessons. You can pay monthly or yearly in Google Play, and usage limits apply.
Sources
Facts and definitions were checked against these public sources. Lesson text and examples were written independently; no text was adapted.
- Machine Learning Glossary
- Rule-based system
- Introduction to Machine Learning: What is Machine Learning?
- Machine learning
- Large language model
- Machine Learning Crash Course: LLMs: What's a large language model?
- Machine Learning Crash Course: Introduction to Large Language Models
- Softmax function
- Hallucination (artificial intelligence)
- Fact-checking
- Primary source
Related courses
- Thinking
Critical Thinking: Arguments and Fallacies
How to read an everyday argument: find its conclusion and the reasons behind it, judge whether those reasons really support it, and spot six common fallacies in ordinary talk.
3 lessons, 18 steps - Programming
Python from Zero
For adults who have never programmed: read short Python programs, work out what they print and choose fixes for common mistakes with variables, if / elif / else and loops. You type or pick each answer; the app does not run code.
3 lessons, 18 steps - Data literacy
Data & Numbers in the News
Read the numbers behind the headlines: percentage points, relative and absolute risk, averages and claims about cause.
3 lessons, 18 steps - Computers Open textbook
Introduction to Computers
Built automatically from early chapters of the open textbook "Introduction to Computers" on Wikiversity: computer types, hardware and software, the processor and future peripherals. Not reviewed by a person.
From Wikiversity, 3 lessons
