Singapore · 481 occupations · public government data
Which Singapore jobs will AI reshape first?
Every occupation in Singapore's resident workforce, scored 0–10 for how much of its
day-to-day work overlaps with what current AI can already do — joined with median wages,
estimated employment, and four years of hiring trends.
—
resident workers mapped
—
work in highly exposed roles (score 7+)
—
of annual wages sit in those roles
A high score does not mean the job disappears — it means the daily work changes.
Some exposed occupations are growing fastest of all.
Check your own job, then scroll the story.
Start here
How exposed is your job?
Search any of the 481 occupations from Singapore's official SSOC 2020 classification. Once you pick one, your dot follows you through every chart below.
The story
One picture, four chapters
Every dot below is one of the 481 occupations. Keep scrolling — the picture rearranges itself.
481 occupations, one dot each
From software developer to bus driver to pre-school teacher —
every common occupation in the resident workforce, colored by AI exposure:
cool = insulated, hot = exposed.
Almost half the workforce sits in the hot zone
Stacked by score, the picture tilts hot. An estimated — residents
— — of the mapped workforce — work in occupations scoring 7 or higher.
Only — are in the cool zone of hands-on, face-to-face work that AI barely touches.
The better-paid the desk, the bigger the overlap
Dots now size to estimated headcount and rise by pay. Past automation hit the factory floor;
AI lands on the office. The median high-exposure occupation pays —
— — the median low-exposure one (—).
Exposure is not destiny
The dots gather into 38 clusters: right = more exposed, up = grew since 2021.
ICT professionals (—) grew — in four years.
General clerks (—) shrank —
— a decline that began before generative AI existed. A high score tells you the work changes, not that demand falls.
The double squeeze
Clerical work was falling before AI. ICT kept climbing
Indexed employment for six notable clusters, 2021 = 100. The most exposed desk jobs split into two fates: clerical clusters were already shedding workers when generative AI arrived — for them AI accelerates an old decline. Highly exposed but expanding fields absorb AI as a tool instead.
Employment indexed to 2021 (=100), via 2023. Line color = worker-weighted AI exposure of the cluster. Source: Labour Force in Singapore, Table D8.
Where it concentrates
The pressure map, cluster by cluster
All 38 occupation clusters, ranked by worker-weighted AI exposure. Tap a cluster to see the occupations inside it.
The full list
Explore all 481 occupations
Methodology
How this was built — and what it can't tell you
The exposure score
Each occupation is scored 0–10 against a calibrated rubric measuring how much of its day-to-day work is digitally mediated and within reach of current AI systems. Anchor points:
SingStat M182081 — employment by industry × occupation group, 2000–2025
Honest limitations
Headcounts are estimates. Employment isn't published at detailed occupation level; group totals are distributed evenly across the occupations in each industry cell. Treat sizes as order-of-magnitude.
Residents only. Citizens and PRs (~2.4M). The ~1.5M work-pass workforce is excluded — construction trades and domestic work are heavily under-represented.
Trends are cluster-level. Employment change is measured for 38 clusters, not for each occupation individually.
Only "common" occupations appear. Wage tables omit roles with small samples — niche or emerging titles are absent.
Scores read the job title, not the worker. Your actual exposure depends on your specific tasks, employer and tools.