Top

Free AI learning, matched to you

In 2026 AI is already part of homework, hiring, health apps and the news, and by 2040 it will be as ordinary as the internet. The people who thrive will not be the ones who memorize tools. They will be the curious ones who can ask good questions, check the answers and decide what is right. The world’s best AI academies now teach for free. This page picks the best steps from them and lines them up by who you are, so you can start today and never pay a cent. If you coach others, the Coaching Kit below gives you a four-week circle, discussion prompts and safety tips. Looking for the full catalogue of providers? Visit the AI Academy Atlas.

What you’ll be able to do

Choose your own road

Pick the path that fits you or your learners, from a family evening to a developer portfolio, with every step free to start.

Earn proof you can show

Collect free certificates and badges from Anthropic, Elements of AI, Kaggle, Hugging Face and more, with honest labels on what is free and what is not.

Use AI with good judgment

Prompt clearly, check the answer, spot bias and know when a person must decide.

Coach a curiosity circle

Run a four-week AI circle with free tools, discussion prompts, a prompt exercise and a myth-vs-fact game.

Keep young learners safe

Know the official minimum ages of the big tools and the habits that protect minors.

Stay current

Follow a short list of free newsletters, podcasts and official blogs so the learning does not stop.

How to read the cards

Each step shows the provider, what you earn, an honest cost label (Free certificate, Free badge, Free course or Free to audit · paid certificate), approximate hours and a level. “Certificate not confirmed” means the provider’s page does not clearly promise one. Links open the official page in a new tab.

(a) Curious beginners & families

No coding, no maths. A calm first month with AI: understand it, use it carefully and try it with your hands. Whole path: about 40 hours, at your own pace (hours are our estimates unless the provider states them).

Step 1 · Elements of AI: Introduction to AI

Provider: University of Helsinki & MinnaLearn. You earn: a certificate from the University of Helsinki when you complete at least 90% of the exercises with 50% of those answered correctly (per the course FAQ). Cost: Free certificate · About 25 hours (estimate) · Level: Beginner

Start Elements of AI Certificate rules

Step 2 · AI Fluency: Framework & Foundations

Provider: Anthropic Academy. You earn: a certificate of completion after the final assessment. Teaches the 4D framework: Delegation, Description, Discernment and Diligence. Cost: Free certificate · About 3 hours (estimate; Anthropic does not state hours) · Level: Beginner

Open the course

Step 3 · Work smarter with AI (AI-3025)

Provider: Microsoft Learn. You earn: an achievement on your Microsoft Learn profile (achievement code available). Plain-language start: what AI is, grounding, agents and checking AI output. Cost: Free course · About 6 to 8 hours (estimate; Microsoft lists it as a 1-day course) · Level: Beginner

Open the course

Step 4 · OpenAI Academy

Provider: OpenAI. You earn: practical know-how for everyday use, plus live events and communities. Certificate not confirmed. Cost: Free course · Self-paced, 2 to 4 hours (estimate) · Level: Beginner

Visit OpenAI Academy All courses

Step 5 · Train your own model with Teachable Machine

Provider: Google. You earn: a hands-on feel for how machines learn from examples (no certificate). Runs in the browser with a webcam or microphone, no code. Cost: Free course · About 1 hour (estimate) · Level: Beginner

Try Teachable Machine

(b) Kids & teens

Hands-on, age-matched and adult-guided. Learners under 18 should use these with a coach, parent or teacher (see the safety tips in the Coaching Kit below). Most of these are classroom resources, so none of them confirm a badge or certificate for learners. Whole path: about 30 hours across a term (hours are our estimates unless the provider states them).

Step 1 · How AI Works (video series)

Provider: Code.org (CodeAI). You earn: a clear picture of how AI learns, in short classroom-friendly videos (grades 6 to 12). No certificate confirmed. Cost: Free course · About 1.5 hours (estimate) · Level: Beginner

Start the series

Step 2 · AI for Oceans

Provider: Code.org. You earn: train a model to sort fish from trash and see how training data shapes results (grades 3 to 12). No certificate confirmed. Cost: Free course · About 1 hour (estimate) · Level: Beginner

Start AI for Oceans

Step 3 · Dance Party: AI Edition

Provider: Code.org. You earn: a playful lesson on AI and creativity for younger learners (grades 3 to 8). No certificate confirmed. Cost: Free course · About 1 hour (estimate) · Level: Beginner

Start Dance Party

Step 4 · Experience AI lessons

Provider: Raspberry Pi Foundation & Google DeepMind. You earn: ready-to-run lessons on AI, machine learning and AI safety for ages 8 to 16 (lesson plans, slides, worksheets, videos). No certificate confirmed. Cost: Free course · About 4 to 6 hours across units (estimate) · Level: Beginner

Browse the units About Experience AI

Step 5 · Day of AI curriculum

Provider: MIT RAISE. You earn: free, grade-by-grade AI activities (PreK to 12), plus a family resources page. No certificate confirmed. Cost: Free course · About 3 to 5 hours (estimate) · Level: Beginner

Open the curriculum Family resources

Step 6 · Artificial Intelligence Discoveries (grades 6 to 8) or Foundations (grades 9 to 12)

Provider: Code.org. You earn: a full free course for a school year or a coaching term. Certificate not confirmed. Cost: Free course · 20+ hours (estimate) · Level: Intermediate

AI Discoveries (6-8) AI Foundations (9-12)

(c) Educators & Curiosity Coaches

Build your own fluency first, then your classroom or circle practice. Pair this path with the Coaching Kit below. Whole path: about 18 hours (hours are our estimates unless the provider states them).

Step 1 · Understanding AI for educators

Provider: Experience AI (Raspberry Pi Foundation & Google DeepMind). You earn: confidence to teach and talk about AI with young people. Free online course for teachers, parents and club volunteers. Certificate not confirmed. Cost: Free course · About 4 hours (estimate) · Level: Beginner

Find the educator course

Step 2 · AI Fluency: Framework & Foundations

Provider: Anthropic Academy. You earn: a certificate of completion. This is the prerequisite for the educator course next. Cost: Free certificate · About 3 hours (estimate) · Level: Beginner

Open the course

Step 3 · AI Fluency for educators

Provider: Anthropic Academy. You earn: a certificate of completion after the final assessment. Applies AI Fluency to teaching practice and institutional strategy. Needs the 4D framework from step 2. Cost: Free certificate · About 3 hours (estimate) · Level: Intermediate

Open the course pK-12 path

Step 4 · AI 101 self-paced professional learning

Provider: Code.org. You earn: classroom-ready understanding of AI for K-12 teachers. Code.org describes it as free for any teacher. Certificate not confirmed. (Code.org's page blocks automated checks, so we confirmed it from the CodeAI hub.) Cost: Free course · About 3 hours (estimate) · Level: Beginner

Open AI 101 CodeAI hub

Step 5 · Day of AI professional development

Provider: MIT RAISE. You earn: free teacher training that goes with the Day of AI curriculum. Certificate not confirmed. Cost: Free course · About 3 hours (estimate) · Level: Beginner

See the training

Step 6 · UNESCO AI competency framework for teachers

Provider: UNESCO. You earn: a shared vocabulary for what teachers should know and do with AI (a reference document, no certificate). Cost: Free course · About 2 hours of reading (estimate) · Level: Intermediate

Read the framework TeachAI toolkit

(d) Business & product people

From “what is this?” to leading a team that works with AI. Product managers should finish at our own PILOT track. Whole path: about 20 hours before PILOT (hours are our estimates unless the provider states them).

Step 1 · Claude 101

Provider: Anthropic Academy. You earn: a certificate of completion. Everyday AI for work tasks and core features. Cost: Free certificate · About 2 hours (estimate) · Level: Beginner

Open the course

Step 2 · AI Fluency: Framework & Foundations

Provider: Anthropic Academy. You earn: a certificate of completion and the 4D habits of mind for delegating to AI and checking it. Cost: Free certificate · About 3 hours (estimate) · Level: Beginner

Open the course

Step 3 · Work smarter with AI (AI-3025)

Provider: Microsoft Learn. You earn: a Microsoft Learn achievement for using AI in everyday business work. Cost: Free course · About 6 to 8 hours (estimate) · Level: Beginner

Open the course

Step 4 · Drive AI transformation in your organization (AB-731T00)

Provider: Microsoft Learn. You earn: strategy skills for spotting high-impact AI uses and championing responsible AI. The course is free; the related Microsoft Certified: AI Transformation Leader credential is a separate paid exam. Cost: Free course · About 6 to 8 hours (estimate; listed as a 1-day course) · Level: Beginner

Open the course

Step 5 · Apply AI at Work

Provider: OpenAI Academy. You earn: repeatable AI workflows and prompting habits for everyday tasks. Certificate not confirmed. Cost: Free course · About 3 hours (estimate) · Level: Beginner

Open the course

Step 6 · PILOT: Agentic Product Management track

Provider: Ignite Curiosity. You earn: our 12-week track for product managers, built from free certifications plus real GitLab practice. Cost: Free course · About 12 weeks, at your own pace · Level: Intermediate

Open the PILOT track

(e) Builders & developers

Agents, APIs and retrieval. You will need Python. Keep a repo of your builds: that is your real portfolio. Whole path: about 45 hours (hours are our estimates unless the provider states them).

Step 1 · Building with the Claude API

Provider: Anthropic Academy. You earn: the full path from first API call to tool use and agent architectures. Needs Python and basic JSON. A final assessment is included; certificate not explicitly confirmed. Cost: Free course · About 8 hours (estimate) · Level: Intermediate

Open the course

Step 2 · Introduction to Model Context Protocol

Provider: Anthropic Academy. You earn: you build MCP servers and clients in Python (tools, resources, prompts). Certificate not explicitly confirmed. Cost: Free course · About 3 hours (estimate) · Level: Intermediate

Open the course

Step 3 · AI Agents Course

Provider: Hugging Face. You earn: a Fundamentals certificate (finish Unit 1) or a Certificate of Completion (Unit 1, one use-case assignment and the final challenge). Free, and you can also audit with no certificate. Needs basic Python. Cost: Free certificate · About 3 to 4 hours per week (Hugging Face), 5 or more weeks · Level: Intermediate

Start the course

Step 4 · Generative AI for Beginners, including lesson 15: RAG and vector databases

Provider: Microsoft (open-source curriculum). You earn: 21 free lessons with code. Lesson 15 builds a retrieval-augmented generation (RAG) app. No certificate; the GitHub repo is the course. Cost: Free course · About 12 hours for the lessons you pick (estimate) · Level: Intermediate

Open the lessons RAG lesson

Step 5 · AI Agents for Beginners

Provider: Microsoft (open-source curriculum). You earn: 18 free lessons on building AI agents. No certificate. Cost: Free course · About 10 hours (estimate) · Level: Intermediate

Open the lessons

Step 6 · Agentic AI

Provider: DeepLearning.AI (Andrew Ng). You earn: 9 hours 55 minutes of video you can audit for free. Quizzes, labs and certificates need a paid Pro membership, so we count it as audit-only here. Needs intermediate Python. Cost: Free to audit · paid certificate (Pro membership, $25/mo billed annually or $30/mo billed monthly) · About 10 hours (listed by DeepLearning.AI) · Level: Intermediate

Audit the course

(f) Data & ML engineers

Python in, working models out. Shorter Kaggle notebooks first, then the deeper courses. Whole path: about 130 hours across a few months (hours are our estimates unless the provider states them).

Step 1 · Intro to Machine Learning

Provider: Kaggle Learn. You earn: hands-on notebooks and a Kaggle certificate of completion (Kaggle shows free certificates on its Learn pages; confirm on the course page). Needs basic Python and Pandas. Cost: Free certificate · About 3 hours (estimate) · Level: Beginner

Open the course Pandas first

Step 2 · Intermediate Machine Learning

Provider: Kaggle Learn. You earn: missing values, pipelines, cross-validation and leakage. Kaggle certificate (confirm on page). Cost: Free certificate · About 4 hours (estimate) · Level: Intermediate

Open the course

Step 3 · Machine Learning Crash Course

Provider: Google for Developers. You earn: 12 modules in four groups (ML models, data, advanced ML models, real-world ML). Certificate not confirmed. Cost: Free course · About 15 hours (estimate) · Level: Intermediate

Open the course

Step 4 · Practical Deep Learning for Coders

Provider: fast.ai. You earn: a code-first path to training and deploying real models. No certificate. Cost: Free course · About 30 hours (estimate) · Level: Intermediate

Start with lesson 1

Step 5 · LLM Course

Provider: Hugging Face. You earn: 12 chapters on transformers, fine-tuning, datasets, tokenizers and sharing models. There is currently no certificate. Needs good Python. Cost: Free course · About 30 hours (estimate) · Level: Advanced

Open the course

Step 6 · CS50’s Introduction to Artificial Intelligence with Python

Provider: Harvard CS50. You earn: search, knowledge, uncertainty, optimization, learning, neural networks and language, in seven weeks of material. Free on OpenCourseWare; the edX verified and professional certificates are paid (see the official page for current prices). Needs CS50x or about a year of Python. Cost: Free to audit · paid certificate (price on official page) · About 50 hours (estimate) · Level: Advanced

Open CS50 AI

(g) Leaders & responsible AI

For board members, directors, coaches and anyone who signs off on AI use. Less typing, more judgment. Whole path: about 22 hours (hours are our estimates unless the provider states them).

Step 1 · Intro to AI Ethics

Provider: Kaggle Learn. You earn: short, practical lessons on bias, explainability and human-centered design. Kaggle certificate (confirm on page). Cost: Free certificate · About 3 hours (estimate) · Level: Beginner

Open the course

Step 2 · AI Fluency: Framework & Foundations

Provider: Anthropic Academy. You earn: a certificate of completion. The “Diligence” part of the 4D framework is about responsible use and accountability. Cost: Free certificate · About 3 hours (estimate) · Level: Beginner

Open the course

Step 3 · Drive AI transformation in your organization (AB-731T00)

Provider: Microsoft Learn. You earn: a leadership plan for AI adoption with responsible AI practices built in. Free course; the related credential has a paid exam. Cost: Free course · About 6 to 8 hours (estimate) · Level: Beginner

Open the course Learning path

Step 4 · NIST AI Risk Management Framework

Provider: NIST. You earn: the U.S. voluntary framework for managing AI risk, with a free Playbook. A reference, not a course (no certificate). Cost: Free course · About 4 hours of reading (estimate) · Level: Intermediate

Read the framework AI RMF Playbook

Step 5 · Recommendation on the Ethics of Artificial Intelligence

Provider: UNESCO. You earn: the first global standard on AI ethics, a good anchor for discussion (a reference, no certificate). Cost: Free course · About 2 hours of reading (estimate) · Level: Intermediate

Read the recommendation

Step 6 · Google’s responsible AI practices and OECD.AI

Provider: Google AI & OECD. You earn: how a large builder describes its practices, next to an independent policy observatory (references, no certificate). Cost: Free course · About 2 hours (estimate) · Level: Intermediate

Google practices OECD.AI

Later, when sponsored

None of these are part of the $0 paths. Prices change, so check each official page before anyone pays.

Google AI Essentials (Coursera)

Provider: Google, on Coursera. You earn: a Google certificate. Under 5 hours, 5 modules. Cost: Paid certificate: $49 USD per month after a 7-day free trial in the U.S. and Canada (other countries may be lower; check Coursera). About 5 hours · Level: Beginner

See the official page (paid)

Microsoft Certified: Azure AI Fundamentals

Provider: Microsoft. You earn: a Microsoft certification after passing an exam; the free Microsoft Learn course “Introduction to AI in Azure” is the matching training. Cost: Paid exam, price varies by country (see the official page). Level: Beginner

Certification details (paid exam) Free matching course

DeepLearning.AI Pro membership

Provider: DeepLearning.AI. You earn: quizzes, labs and certificates on its courses. Cost: $25/mo billed annually or $30/mo billed monthly (as listed on the Agentic AI course page). Level: Intermediate

See the official page (paid)

Free hands-on playgrounds

Teachable Machine

Train an image, sound or pose model in your browser. Open it.

Quick, Draw!

Doodle and watch a neural network guess. Play.

TensorFlow Playground

Watch a small neural network learn, with sliders. Open it.

Kaggle and Colab notebooks

Run Python in the browser. Kaggle Learn · Google Colab.

Coaching Kit

Everything a Curiosity Coach needs to turn the paths above into a living, human learning experience. Nothing here needs a paid tool.

Run a 4-week AI curiosity circle

Six to ten learners, 60 minutes a week. The coach facilitates; the learners do the thinking. For learners under 18, the coach holds the account and drives any chatbot on screen.

Week 1 · Notice AI

Warm-up: where did you meet AI this week? Then train a model together with Teachable Machine and play Quick, Draw!. Ask: how did the computer learn, and what did it get wrong?

Week 2 · Prompt and check

Run the prompt-practice exercise below on one shared screen. Compare three versions of a prompt, then fact-check the answer against a trusted source.

Week 3 · Fair and safe

Do AI for Oceans to see how training data shapes results, then use two or three of the discussion prompts below. Finish with the myth-vs-fact game.

Week 4 · Make and showcase

Each learner shares one thing: a small project, a before-and-after prompt, or a question they still have. Plan their next free course from the paths above and show them how to display a badge.

Discussion prompts: ethics, bias and safety

  • An AI learned from photos that mostly showed one kind of person. Who might it treat unfairly? How could we find out?
  • If an AI gives a wrong answer with total confidence, who is responsible for the mistake: the maker, the user, or both?
  • What should a person always decide, even when an AI is available to help (a grade, a medical choice, a hiring decision)? Why?
  • Is it fair to use AI for homework? Where is the line between help and cheating in your class or workplace?
  • What personal information would you never type into a chatbot? What could happen to it?
  • How can you tell whether an image, video or voice is real? What would you do if someone shared a fake of you?
  • Who is missing from the data an AI learned from, and who benefits when AI works well?
  • When is it kind and honest to say “an AI helped me with this”?

Deeper reading for coaches: the UNESCO ethics recommendation, the NIST AI Risk Management Framework and the UNICEF policy guidance on AI for children.

Prompt-practice exercise

Role, Task, Context, Format, Check (our five-part prompt)

Pick a real task, such as planning a family picnic or explaining photosynthesis to a ten-year-old. Work in four rounds and write down what changes.

  1. Vague prompt: type only the topic. Read the answer together.
  2. Add Role and Task: “You are a friendly science tutor. Explain…”
  3. Add Context and Format: who it is for, how long, a table or a list.
  4. Add Check: ask it to list what it is unsure about, then verify two facts yourself with a trusted source.

Debrief: which round helped most, and what did the AI still get wrong? Never paste private or confidential information. For a guided version of the same idea, see the 4D framework (Delegation, Description, Discernment, Diligence) in Anthropic’s AI Fluency course.

AI myth vs fact

Read each myth aloud. Learners vote “myth” or “fact,” then the coach reveals the answer and the group explains why.

Myth: AI understands things like a person.

Fact: Today’s AI finds patterns in data and predicts likely outputs. It can sound human without having human understanding or experience.

Myth: If AI says it, it must be true.

Fact: AI can state wrong things confidently. Always check important facts with a trusted source.

Myth: AI is neutral, so it cannot be biased.

Fact: AI learns from human-made data and design choices, so it can repeat or amplify unfairness.

Myth: You need to be a coder to learn AI.

Fact: Many free courses above need no code at all. Coding helps for building, not for understanding or using AI well.

Fact: How you ask changes what you get.

Clear context, a goal and a format usually give better answers. That is why prompt practice is a real skill.

Fact: People stay accountable for AI-assisted work.

A tool can help draft, but a person owns the decision and the result.

Showcase your badges

  • Collect proof as you go: save each certificate PDF or badge link in one folder named for the course.
  • Credly badges (for example IBM SkillsBuild): accept the badge in your Credly account, then use its share button to post it to LinkedIn or copy the public link.
  • On LinkedIn: add each credential under the “Licenses & certifications” section of your profile and paste the credential URL. Add the skill it taught to your Skills list.
  • Say what you did: one sentence per badge, such as “Completed Anthropic’s AI Fluency course and applied it to plan a community workshop.”
  • Under 18? Social networks have their own minimum ages, so ask a parent or guardian and check the network’s rules. A printed certificate or a simple portfolio page the family controls works well too.
  • Be honest about the label: a certificate of completion is not an exam credential. Say which one you earned.

Safety tips for minors using AI tools

Minimum ages are set by each company and change over time, so re-check the official page before every new term. We list what each official page says today.

ToolOfficial minimum ageSource
Claude (Anthropic)18 or older to create and use an account.Claude Help Center
ChatGPT (OpenAI)Not meant for children under 13; ages 13 to 18 need parental consent. (OpenAI’s help page blocks automated checks, so we confirmed this through search of the help center.)OpenAI Help Center
Gemini (Google)13 (or the applicable age in your country) or over for a personal or school Google Account. A Family Link account for a child under 13 cannot use the Gemini web app. Work accounts must be 18 or over.Gemini Apps Help
Microsoft CopilotAt least 13, and higher in some countries depending on local rules. Microsoft notes a newer Copilot app may differ, so read its current article.Microsoft Support
Khanmigo (Khan Academy)18 or older to sign up directly; under 18 needs a parent or guardian to add them. Learner access is paid; teacher tools are free.Khanmigo
  • Coach-led for the youngest: if a tool’s minimum age is not met, an adult runs the screen and the child only watches or suggests ideas. Use the free browser tools in Week 1 (Teachable Machine, Quick, Draw!) and Code.org lessons, which are built for classrooms.
  • No personal details: no full names, addresses, school names, photos of faces, passwords or private family information in any prompt.
  • Check everything: teach the habit “AI said it, now I verify it.” Do not rely on a chatbot for health, safety or emotional crises. Talk to a trusted adult.
  • Chatbots are not friends: keep a human in every important conversation, and tell a trusted adult if something an AI says feels strange or upsetting.
  • Deepfakes and kindness: never make or share a fake image, voice or video of a real person. Talk about what to do if one appears.
  • Make a family or circle agreement: which tools, when, and who sits nearby. Review it each term.
  • More guidance: Common Sense Media on AI, UNICEF policy guidance on AI for children and National Cybersecurity Alliance.

Stay current

AI moves quickly. Ten minutes a week beats a binge once a year. All of these are free to read or listen to (some newsletters also offer paid tiers you do not need).

Newsletters
Podcasts
Events and communities
Official blogs and reports

Free resource library

Learning Paths FAQ

Every step in the seven paths is free to start. Where a certificate costs money (DeepLearning.AI, CS50 on edX, Google AI Essentials) we say so on the card, and those paid items sit in the “Later, when sponsored” block or are marked free to audit.
Not sure? Take path (a). It needs no coding and gives you a certificate in the first week. Then choose the path for your role.
We only promise what an official page states. If a provider does not clearly say a certificate or badge exists, we say so rather than guess.
No. They are our estimates unless the provider lists hours. Plan for more time if you are new to the topic.
It depends on age and the tool. See the minimum-age table in the Coaching Kit. For younger learners, a coach or parent should lead and use classroom-built tools such as Code.org lessons.
Use Credly or the provider’s share button, add the credential on LinkedIn if you are of age, and keep a folder of proofs. See “Showcase your badges.”

Track your progress

Make a simple checklist with one line per step. Tick a step only when you have finished the course or the final assessment, and paste the certificate or badge link next to it.

Pick one step and begin today. Curiosity grows when you use it.