Learn AI · Level 1 · Foundations · step 13 of 13
The best free AI courses, from beginner to expert
Reading about AI is a great start; trying it is how it sticks. These nine courses are all free, come from organisations that know AI best, and are sorted from no-coding-needed to building your own models. Each one fits a level of our Learn AI path.
No coding needed
Start here if you have never written code. Both take a few hours, spread over as many days as you like.
Best first course
Beginner1. Elements of AI by University of Helsinki and MinnaLearn
The best first course on AI for non-technical people. Short chapters with exercises explain what AI is, how machine learning and neural networks work, and what AI means for society, with no maths or coding needed. Over 2 million people have signed up.
- No maths or programming needed
- Clear, practical exercises at the end of each chapter
- Available in many languages
- Go at your own pace
Watch out: The follow-on course, Building AI, is more technical and recommends basic Python.
Price: FreeOn: Web
Best for using AI well at work
Beginner2. AI Fluency: Framework and Foundations by Anthropic
A short course on working well with AI chatbots in any job: when to hand a task to AI, how to describe what you want, how to judge what comes back, and how to use it responsibly. It's the starting point of Anthropic's free course catalogue and ends with a certificate.
- Useful for any role, no technical background needed
- A simple framework you can apply to any chatbot
- Free certificate on completion
Watch out: It's made by Anthropic, so the examples use its Claude chatbot, though the ideas apply to any.
Price: Free, with an email sign-upOn: Web
Learn how it works
Short, hands-on courses where you build and test real models. A little Python helps.
Best hands-on start
Beginner3. Intro to Machine Learning by Kaggle
A hands-on introduction in about three hours: how models work, building your first model, measuring how good it is, and avoiding underfitting and overfitting. Each short lesson is followed by an exercise you run in the browser.
- Learn by doing, with code that runs in your browser
- About three hours from start to finish
- Free certificate on completion
Watch out: Uses Python. If you've never coded, take Kaggle's free Intro to Programming course first.
Price: Free, with a Kaggle accountOn: Web
Best on fairness and bias
Beginner4. Intro to AI Ethics by Kaggle
A practical course, about four hours long, on designing AI that treats people fairly: human-centred design, how bias creeps into AI systems, different definitions of fairness, and "model cards" that explain what a model can and can't do.
- Real scenarios, such as a credit-card approval model
- Short, focused lessons
- Free certificate on completion
Watch out: The exercises use some Python, but the lessons are readable without coding.
Price: Free, with a Kaggle accountOn: Web
Best all-round foundation
Intermediate5. Machine Learning Crash Course by Google
Google's well-known introduction to machine learning, with animated videos, interactive charts and coding exercises. It covers how models learn, how to prepare data, neural networks, language models and responsible AI, in self-paced modules of roughly 13 hours in all.
- Interactive visualisations make the maths intuitive
- Exercises run in the browser on Google Colab
- Take it in order, or pick the modules you need
Watch out: Expects basic algebra and statistics and some Python. Newcomers should start with Elements of AI.
Price: FreeOn: Web
Build with AI
For people who want to make apps, agents and tools with today's models.
Best for building AI apps
Intermediate6. Generative AI for Beginners by Microsoft
21 lessons on building apps with generative AI: prompts, chat and search apps, RAG, image generation, agents, fine-tuning and responsible use. Each lesson stands alone, with code in Python and TypeScript, published openly on GitHub.
- Covers the whole toolkit for building AI apps
- Start with any lesson
- Very popular and actively maintained
Watch out: Some lessons call paid AI services (Azure OpenAI or the OpenAI API), so running every example may cost a little.
Price: Free course; some examples use paid AI servicesOn: Web, GitHub
Best on AI agents
Intermediate7. 5-Day AI Agents Intensive by Google and Kaggle
Google's course on AI agents, first run live in November 2025 and now free to take at your own pace. Each day combines background reading and podcasts with hands-on coding labs on models, tools, memory, orchestration and evaluation, and how agents become production-ready.
- Built by Google ML researchers and engineers
- Theory and hands-on labs side by side
- Covers testing and evaluating agents, which most courses skip
Watch out: Needs free Kaggle and Google AI Studio accounts, and some coding. Optional cloud deployment may need a Google Cloud billing account.
Price: FreeOn: Web
Best for open models
Advanced8. LLM Course by Hugging Face
A free, ad-free course in 12 chapters on how large language models work and how to use, fine-tune and share them with the open-source Hugging Face tools. It is the standard hands-on route into working with open models.
- From transformers basics to fine-tuning your own model
- Uses the tools most AI teams actually work with
- Free and without ads
Watch out: Needs good Python, and a basic deep learning course first is recommended. There is no certificate at the moment.
Price: FreeOn: Web
Go deep
The course many AI engineers started with.
Best deep learning course
Intermediate9. Practical Deep Learning for Coders by fast.ai
The classic "top-down" deep learning course: you build working models for images, text and data from the first lesson, then learn how they work underneath. Nine video lessons of about 90 minutes, based on a free book, using PyTorch and Hugging Face tools.
- You build real models from lesson one
- Teaches the maths it needs, school-level maths is enough
- Uses free cloud notebooks, no special computer needed
Watch out: Aimed at people with about a year of coding experience, ideally in Python.
Price: FreeOn: Web, YouTube
How to use this list
- Not technical? Take Elements of AI, then AI Fluency. Together they cover what AI is and how to work with it well.
- Curious how it works? Add Kaggle’s Intro to Machine Learning: in about three hours you’ll have trained a model yourself.
- Want to build things? Microsoft’s Generative AI for Beginners, then the Google agents course.
- Aiming to work in AI? Google’s Machine Learning Crash Course, then fast.ai and the Hugging Face course.
Each course pairs with a level of our Learn AI path, which explains the ideas behind them in plain English. Prefer reading? See the best AI books, or try the free simulators in our AI Playground.
How we chose
We picked courses that are free to take, come from organisations with real AI expertise, are kept up to date, and teach by doing. We checked each against its official page, including what it costs and what it expects you to know already. Free certificates are noted where offered. We review this list every quarter. Know a course we should add? Tell us.