Alison has been running since 2007 and has grown into one of the largest free learning platforms anywhere, with tens of millions of registered learners. The model is straightforward: study for nothing, sit through advertising while you do it, and pay only if you want a certificate or diploma at the end. I have more respect for that than for the free trial that quietly becomes a subscription, because at least you know what the deal is from the start. The AI and machine learning section has grown quickly, covering introductions to artificial intelligence, machine learning fundamentals, generative AI and ChatGPT for work, AI in business, and various vertical applications.
Certificate courses sit around two to three hours and diplomas around eight to ten. That is a sensible amount of time for orienting yourself in an unfamiliar subject, and if you have never thought about what a model actually does then a few hours here will leave you better informed than you were. The problem is what happens next. Very little of this material puts you in front of code, a dataset or a live model.
It explains, it quizzes you on the explanation, and it ends. Learning AI without building something is a bit like learning to swim from a diagram, and the catalogue does not really acknowledge that. Compare it to Kaggle Learn, which is also free, or the free tier of Google's own machine learning material, and Alison comes off looking thin. Quality control is the other issue.
A large slice of the catalogue is licensed from third party publishers or contributed by outside providers, so the standard varies enormously from one course to the next. Some are competently produced. Others feel like a slide deck with narration bolted on, with production values and recency that would embarrass a YouTube channel. Because publication dates are frequently vague, you can finish a generative AI module before realising the tool landscape it describes has moved on considerably.
In a field where a year is a long time, that matters more than it would in bookkeeping or health and safety. Then there is the certificate. Alison markets these as CPD accredited, and technically that claim holds, but CPD accreditation is a much lower bar than most learners assume and does not mean an employer will recognise or value the document. Paying twenty five dollars for a PDF that no hiring manager has heard of is not a scandal, but I would rather see people spend that money on compute credits and build something instead.
If you want proof of skill in AI, a small project you can show and explain will outperform any Alison diploma in an interview, every time. My 3.1 is a free resource doing a real service for people who cannot pay for anything better, marked down for a shallow AI curriculum, uneven quality and a certificate business that trades on learner misunderstanding. Use it to work out whether the subject interests you. When it does, move on quickly to somewhere that makes you build.