The Jobs AI Will Kill First in Malaysia (and the Ones It Can't Touch)

AI does not come for every job equally

Let me be precise, because the panicking version is useless. AI comes hardest and first for one kind of work: routine, repetitive, predictable work where standard inputs produce standard outputs. If most of your day is the same predictable task, you are in the blast radius. If it is messy, human and different every day, you are safer.

Look at the roles already feeling it: data entry, basic customer service and contact centres, administrative and clerical work, bookkeeping, basic content production. McKinsey estimates around 30% of tasks in areas like admin and customer service could be automated by the end of the decade. Some analyses put the automation risk for routine customer service and data entry far higher.

To be fair, the World Economic Forum expects AI to displace around 92 million jobs globally by 2030 while creating about 170 million. On paper, a net gain. The catch is in the fine print: most new jobs are in roles that barely existed a few years ago. Jobs disappear from one group of people and appear for a different group. Which group you are in is the entire question.

Malaysia's specific exposure

For the last two decades, a big part of our strategy was to become the region's back office. We built shared-services hubs in Cyberjaya and KL for the world's biggest companies: customer support, back-office finance and accounting, data processing. We sold ourselves as the place that does this work well and cheaply.

Now compare that with the list of jobs AI eats first. Customer service, back-office finance, data processing. It is almost the same list. We built a meaningful slice of the economy on precisely the routine work AI is automating fastest. That is not a coincidence. That is exposure.

And there is a cruel irony. The data centres going up in Johor, using our land and water, are the engines running this AI. We are hosting the technology that automates the jobs we built our economy on. Exposed from both directions at once. I wrote more about that in why Malaysia hosts the AI revolution but does not get rich from it.

Security and routine used to be the same thing

I am not saying this to scare you. I am saying it because if you sit in one of these roles thinking you are safe because the company is big, foreign and stable, you are reading the situation backwards. Big, foreign and cost-focused is exactly who automates first.

Picture someone in a back-office job at one of these centres: decent salary, air-conditioned office, a famous logo on the payslip. From outside it looks like they made it. But their day is processing the same kinds of requests over and over, the exact pattern AI handles for a fraction of the cost without sleeping. They feel secure, but they are sitting in one of the most automatable seats in the building. The cleaner and better defined a job is, the better it looks on a cost-cutting slide at headquarters.

Security and routine used to be the same thing. They have quietly become opposites.

What is safer, and the uncomfortable pattern

No job is 100% safe, but lower-risk roles share some traits:

The status ladder and the safety ladder have come apart

Notice something uncomfortable: almost none of those safer roles are the ones we were pushed towards for status. In Malaysia we were told to chase the office job, the corporate desk, the clean white-collar path, and to look down a little on hands-on, technical, messy human work. Many of those clean desk jobs are now in the firing line, while the work we were taught to look down on is more defensible. That alone should make you question advice you are still following.

But do not fall into the opposite trap of "safe jobs vs doomed jobs". Even safe jobs are changing, and risky jobs will still exist with fewer people doing them. Hunting for a safe job is playing defence in a game that rewards offence. For a role-by-role view, see our will AI replace my job guides.

The real divide is between people

The gap opening up is not between safe and risky jobs. It is between people who use AI and people who do not, and it is becoming a chasm. Professionals with AI skills are already reported to command salaries up to 56% higher than people in the same role without them. Analysts expect the ability to direct and supervise AI to be as basic as using a spreadsheet by the end of the decade. Not using AI at work will soon look the way not using email looks today.

Picture two people in the same role at the same company on the same pay a year ago. This year, one started using AI to draft, research and handle the repetitive parts, and now gets through work that used to take a small team. The other still does everything by hand. When the company decides who to promote and who to let go, it is not close. And the first person did not need a new degree or certificate. They just refused to keep doing by hand what the machine could do. AI does not replace people. People who use AI replace people who do not.

The move: hand off the routine, build what the machine can't

Stop hunting for a magically safe job. Spend that time learning to use AI in your day-to-day work. The move is the same in every field: make AI do the routine part, say the boring 60%, and spend your human effort on the 40% it cannot do: judgement, relationships, building and deciding.

For some of you that means adding AI skills to the career you have. For others, especially if your job is mostly routine, it means a bigger move: learning to build with and direct AI as a new skill before the routine version of your job thins out. Analysts put the window for the most exposed roles at 12 to 24 months. Not long, but enough if you start now instead of when the redundancy letter arrives.

Concretely: if you are in admin or support today, do not panic and quit. Start this month using AI on the repetitive half of your work, then use the freed hours to learn something the machine cannot do, such as building, analysing or managing. Within a year you are not the person doing the routine task; you are the person directing it, with proof. That is why I rebuilt my own business around AI rather than treating it as a threat to avoid. The people who get crushed are rarely those in the riskiest jobs. They are the ones who saw it coming and waited. If building with AI is the skill you want, our free trial is the cheapest way to start, and the AI-Native Software Development Programme is the full path.