77,999 tech jobs cut in H1 2025 linked directly to AI. Youth unemployment at 10.2%. 35.5% of graduates are underemployed. We wish the picture was less clear than it is.
Our first reaction to,
“AI is going to take your job” was not ours.
We suspect yours was too. The accountant assumes it will happen to the factory worker. The content writer assumes it will happen to the customer service rep. The customer service rep assumes it will happen to the data entry clerk.
The data entry clerk does not know what to think because the data entry job is already quietly gone. AI-driven job displacement does not announce itself. It quietly makes roles redundant over 12 to 24 months while the headline unemployment numbers look fine on paper.
The numbers I found that Malaysia cannot ignore
| Data point | Figure | Source |
|---|---|---|
| Tech jobs directly cut due to AI (H1 2025) | 77,999 globally | Challenger Gray, 2025 |
| Employers planning to reduce headcount via AI | 41% globally | WEF Future of Jobs Report 2025 |
| Full-time job equivalents at risk globally | 300 million | Goldman Sachs 2023, updated 2025 |
| Jobs projected displaced by 2030 (WEF) | 92 million | WEF Future of Jobs Report 2025 |
| New jobs projected created by 2030 (WEF) | 170 million | WEF Future of Jobs Report 2025 |
| AI-skilled workers earn more than peers | 25% more on avg | PwC AI Jobs Barometer 2025 |
| Malaysia youth unemployment (15-24, Q1 2026) | 10.2% | DOSM, March 2026 |
| Malaysian graduates underemployed (Q3 2025) | 35.5% (1.96 million) | DOSM Q3 2025, The Edge |
Which jobs are already being displaced
According to the Goldman Sachs and McKinsey’s research on which task categories have the highest automation risk. Applied to Malaysian job categories, here is what we found:
High displacement risk (already happening)
- Data entry and processing: 38% of tasks automatable. Basic bookkeeping, invoice processing, data migration. Malaysian banks and shared services centres are already deploying robotic process automation (RPA) for these.
- Customer service (first level): 41% automatable. Chatbots handle FAQ, complaint routing, and basic resolution. Call centres in KL and Cyberjaya have already reduced headcount significantly since 2023.
- Basic content writing and translation: AI writes first drafts, generates product descriptions, and translates documents faster and cheaper than entry-level writers. The entry-level content job market has shrunk measurably since 2023.
- Administrative and scheduling: 46% automatable (Goldman Sachs). Calendar management, meeting notes, travel booking, basic correspondence. AI handles these at near-zero marginal cost.
Medium displacement risk (transition underway)
- Junior accounting and finance: Audit sampling, financial modelling, variance analysis increasingly AI-assisted. The junior analyst who used to spend 40 hours on a report now needs to check AI output for 8 hours.
- Legal support and paralegal work: Contract review, case research, document drafting. Malaysian law firms are beginning to use AI for these tasks, reducing the volume of paralegal work needed.
- Basic software testing and QA: Automated testing tools reduce the human QA headcount required for routine test cases.
Lower displacement risk (for now)
- Complex negotiation and relationship management, physical trades (electricians, plumbers, technicians), clinical and patient-facing healthcare roles, and roles requiring contextual human judgment in novel situations.
What I think the transition actually looks like in Malaysia
The WEF data shows 77% of employers plan to upskill staff for working with AI and 47% plan to move affected employees into different roles.
In Malaysia, where 35.5% of graduates are already underemployed in roles below their qualification level, this transition is happening into a labour market that was already struggling to absorb educated workers into high-skill jobs before AI arrived.
HRD Corp scaled training assistance to RM2.62 billion in 2025 and the government has restrategised TVET towards high-growth industries under RMK-13.
But honestly, the gap between how fast AI is being adopted and how fast the workforce is being reskilled is significant.
PwC’s 2025 AI Jobs Barometer found that industries with higher. AI adoption have seen productivity growth four times higher than less AI-intensive sectors.
Workers with demonstrable AI skills earn 25% more than peers without. The divergence is already showing up in salary data.
Honest take for a Malaysian worker in 2026
The WEF net projection is that AI creates more jobs than it destroys by 2030. That sounds reassuring until you read the fine print: the net number conceals painful individual transitions.
The new jobs being created require genuinely different skills than the jobs being displaced. If you are a Malaysian admin executive, customer service officer, or entry-level content writer right now, you should not be waiting for the disruption to arrive at your specific door. The reskilling needs to start before it becomes urgent.
