Elements of AI: A Proper Introduction to How AI Works

I moved onto Elements of AI after spending some time with the University of Helsinki's Python course.

I'd enjoyed their learning environment enough that when I started looking for a proper introductory AI course, going back to Helsinki felt like the obvious choice. It was free, it had a follow-on course, and by then I already trusted the way they taught.

At that stage, I was curious more than knowledgeable. I'd seen what LLMs could do and could already tell they were going to be important, but I didn't really understand what was happening underneath.

Elements of AI gave me that first structured view.

Learning how to think about AI

The course isn't really about building anything.

Instead, it introduces the ideas behind AI: probability, different forms of machine learning, classifiers, neural networks, problem-solving and the wider implications of using these systems.

The original course was launched in 2018 and is designed specifically for people without a programming background. Its six-part structure covers everything from defining AI and problem solving through to machine learning, neural networks and societal implications.

What I liked was the level it pitched those ideas at.

You aren't expected to become an expert in nearest-neighbour classifiers or suddenly understand the mathematics behind neural networks. It gives you enough exposure to start recognising the concepts and understanding roughly where they fit.

That was exactly what I needed.

It feels like education, not just content

One of the strongest parts of the course was the learning experience itself.

There is a mixture of reading, interactive exercises and written responses. Some of the longer answers are peer reviewed, and you also review work submitted by other learners.

That sounds like a small detail, but I found it useful.

You'd read one answer and think someone had understood the subject brilliantly. Then you'd read another and realise you'd probably gone into more depth yourself. It gave you some sense of where your own understanding sat rather than just moving through a series of automated right-or-wrong questions.

That university influence comes through throughout the course.

It isn't trying to teach you one tool or get you to build an app by the end. It's trying to broaden how you think about AI, including the ethical questions surrounding it.

Has a 2018 AI course aged?

Given how quickly AI moves, the fact that Elements of AI first launched in 2018 might sound concerning.

I didn't find that a problem when I took it.

The course focuses mainly on fundamentals rather than whichever model or product is fashionable at the time. Concepts such as probability, supervised and unsupervised learning, classifiers and neural networks don't suddenly become irrelevant because a new LLM has launched.

The University of Helsinki still positions Elements of AI as a course for a broad audience with no previous AI knowledge required. It has since been taken by well over a million learners around the world.

Who I'd recommend it to

I'd recommend Elements of AI to somebody who is relatively new to the subject and wants to understand more than just how to use ChatGPT.

You don't need to be technical. Equally, somebody who is very comfortable with computers could still get plenty from it if they've never really looked at what sits behind AI systems.

What it won't do is teach you how to build an AI application tomorrow.

That's not the point.

It gives you a wider foundation to build on, and for me it did that well enough that finishing the course didn't feel like reaching the end. It made moving straight onto Helsinki's Building AI course feel like the obvious next step.