In 2018, the University of Helsinki and the education company MinnaLearn set themselves a specific goal: teach the basics of artificial intelligence to one percent of the Finnish population. No ads, no fees, no computer science degree required. The course that came out of that goal, Elements of AI, long since passed its target and kept going. Over a million people have now taken it, in many languages, across most of Europe.
Plenty of free courses exist. Very few were built as public infrastructure, and fewer still stayed free all the way through. That is what makes this one worth understanding, whether you are about to take it or just wondering why it keeps coming up in every list of free AI courses.
An origin story that worked
The course was a deliberate national project: if a small country wanted to be ready for AI, its citizens would need to understand it, and understanding should not depend on income or a university enrollment. The EU picked up the idea and helped spread the course across member states, with translations into many official EU languages — which is rare, and matters, because most free AI education exists only in English.
What makes it different
Most online courses are built around video. Elements of AI is built around text and interactive exercises. You read a plain-language explanation, then you test your understanding by doing something: estimating probabilities, classifying examples, working through how a simple learning system behaves. The exercises do more teaching than the reading does.
Three things follow from that design:
- It is genuinely free, including the certificate. This is the rarest part. Almost every major platform gives away the content and charges for the paper. Elements of AI does not.
- It takes about 30 hours across six chapters, so it is a real course, not a tasting menu.
- The math stops at arithmetic. The course explains the ideas behind machine learning without requiring anything beyond school-level comfort with numbers.
Who it suits
Complete beginners, first of all — people who feel the field passed them by and want to catch up without embarrassment. Teachers, because the material is clear enough to borrow for a classroom. Curious teenagers, because nothing in it assumes a job or a career; it is just an explanation of how a technology works. Anyone with an hour here and there can work through it at their own pace, since progress saves as you go. If you already work with machine learning, you will find it too basic. That is a feature of the audience, not a flaw in the course.
How to take it
The course rewards a simple schedule: one chapter a week for six weeks keeps the thirty hours from collapsing into a resolution you abandon by February. Do the exercises rather than reading past them — on this course more than most, the exercises carry the teaching, and skimming them returns a fraction of the value. Chapters build on each other, so take them in order the first time.
And claim the certificate even if you never show it to anyone. Finishing is the habit this blog keeps coming back to: the courses that change something are the ones that get finished.
The honest limits
Two. First, parts of the course predate the generative AI wave, and the sections on language models and image generators feel dated next to what the tools can now do. The fundamentals still hold, but do not come here to understand what ChatGPT is.
Second, it is conceptual, not practical. You will understand how systems learn; you will not finish with prompts, tools, or workplace workflows. Treat it as the foundations course and pair it with something hands-on if your goal is using AI at work. Thirty hours is a real investment, so it is worth knowing in advance which half of the subject you are getting.
Still the right first course
For most people who want to understand AI rather than feign understanding in meetings, Elements of AI remains the cleanest starting point available. Thirty hours, no cost anywhere in it, and a certificate you can actually show — from a university, in your own language, with no subscription waiting at the end.
No comments yet