Malaysia’s Ministry of Digital is paying for 100,000 Malaysians to take online courses on Coursera. Its aim is to prepare the workforce for AI. The courses people chose first are a useful clue about what that preparation looks like.
The programme runs through the ministry’s Rakyat Digital portal and supports the country’s AI Nation 2030 plan. Learners can reach more than 16,000 courses and professional certificate programmes. The figures here come from the ministry’s own statement, reported on 8 October.
What the early numbers show
In the first three weeks, 12,900 learners enrolled in 43,000 courses. That is about 3.3 courses each. The ministry named the most popular four.
The first is “Understanding Digital Transformation”, a course built for the programme by the ministry’s delivery agency, MyDIGITAL Corporation, using Coursera’s course-building tool. The others are Google’s “AI Fundamentals”, “Foundations: Data, Data, Everywhere” and “Foundations of Project Management”.
Only one of the four is about AI by name. Three are about how organisations change, how data is used, and how projects are run.
What this does and does not mean
Treat it carefully. These are enrolments, not completions. A person can enrol in five courses on a Sunday evening and finish none, and 3.3 courses each may be browsing as much as study.
The ministry reports no results: no skills gained, no jobs changed, no productivity measured. Nobody should call the programme a success on this evidence, and this post does not.
The courses on the list may also be popular for reasons unrelated to demand. Programme promotion, course placement on the portal and the fact that a custom course is easy to find all push enrolments in particular directions.
Why it is still worth noticing
Even with those caveats, the list is a rare piece of evidence. It shows what a large number of working adults picked when they were offered almost anything for free.
They did not all pick a technical AI course. They picked a mix of one introduction to AI, and three courses on how work is organised. That fits what many managers say privately: the hard part of using AI is rarely the tool.
Using AI well at work depends on knowing your process, your data and your decisions. Someone who cannot describe how a piece of work moves from request to result will not be helped by an assistant, however good. They will automate confusion.
A familiar pattern
Consider a finance team given a new AI assistant for reporting. The first weeks go to the tool: which button does what, how to phrase a request. Then someone asks the assistant to reconcile two spreadsheets and the result is plausible but wrong.
The tool was not at fault. The team had never written down how the reconciliation is meant to work, so nobody could say whether the answer was right. The learning that would have helped was about the process, and it was available all along.
Courses on project management and data foundations teach exactly that habit. They ask how the work is broken into steps, what each step needs, and how anyone would know it had gone well. Those are the questions an AI assistant cannot answer for you.
Digital transformation courses do something similar at a larger scale. They put the technology in the context of a business that has to change how it operates. For a manager, that is closer to the real job than a prompt-writing tutorial.
The government’s choice
There is a policy point too. The ministry chose breadth of access over a specialist track. It bought a library, not a bootcamp, and left people to find their own route.
That has a cost. Without guidance, many learners will pick whatever looks interesting and drift. It also has an advantage, since managers, administrators and analysts are not forced into a course designed for engineers.
For anyone responsible for a team’s learning, the lesson is not to copy a national scheme. It is to notice what a scheme chose to leave out.
What to do this month
Start with the work, not the tool. Ask each person to write down, in five steps, how one regular task gets done. Include where it comes from, who checks it and where it goes next. It is one of the most useful pieces of AI preparation there is, and it needs no software.
Pick one foundation course for the team. Choose something on data, process or project management, not a tool tutorial. Ask people to complete it, not merely enrol.
Measure finishing. Track who completes, not who signs up. If you offer a hundred places and ten people finish, you have learned something about time and priority, not about interest.
Give time inside the working week. The Singapore survey this week found that lack of time was one of the top reasons people did not train. A course taken on a Sunday is a course that gets dropped.
Revisit in a quarter. Ask what people changed in their work as a result. One concrete change is worth more than a certificate.
The takeaway
A government spending public money on a library of courses found that people reached for foundations. That is not proof, but it agrees with experience. Before a team asks what AI can do for its work, it helps to be able to say clearly what the work is.
Source: Ministry of Digital, Malaysia, statement reported by TNGlobal on 8 October 2026. Enrolment figures are the ministry’s own and describe the first three weeks of the programme.