Healthcare assistant (e.g. therapy aide/assistant)
AI displacement risk
14%
Range 11.15–15.96%
Healthcare assistant (e.g. therapy aide/assistant) has 34% AI task overlap and 50% human bottleneck protection — lower risk than 51% of occupations in the live market. Current AI capabilities have limited overlap with core tasks.
Service & Sales Workers·SGD 2,432/mo (2,084–2,870)·~2.7K workers in SG·Updated 2026-06-11
Structural pressure, not a prediction of job loss. Displacement tends to arrive through slower hiring, wage compression and role redesign before layoffs.
Why This Score
34% of tasks overlap with current AI
50% human advantage from judgment & presence
40% demand buffer from the local labour market
AI usage 11pp above theoretical exposure
These factors interact with each other — the final score is not a simple sum of these bars.
How much AI overlaps with this job's tasks, offset by human advantages and local demand. Score stability: watch. How this works
Tasks AI can handle
With 34% AI task overlap (based on Felten AIOE, Anthropic Economic Index, and Eloundou GPT exposure), the Healthcare assistant (e.g. therapy aide/assistant) tasks most exposed include: reservation management, menu recommendations, order processing, loyalty program tracking, and basic customer query handling via chatbots.
- • Specify inputs accessed by the system and plan the distribution and use of the results.
- • Test, maintain, and monitor computer programs and systems, including coordinating the installation of computer programs and systems.
- • Use object-oriented programming languages, as well as client and server applications development processes and multimedia and Internet technology.
O*NET tasks for this occupation with the most observed AI usage (Anthropic task data).
What AI can't do here
At 50% human bottleneck protection, the tasks that remain hardest to automate for Healthcare assistant (e.g. therapy aide/assistant) include: genuine hospitality and warmth, reading customer moods, handling complaints gracefully, creating memorable experiences, and adapting service to cultural expectations.
Main insulation channels: Non-routine work + Accountability for others — the work-context dimensions behind this occupation's human bottleneck.
- • Determine computer software or hardware needed to set up or alter systems.
Highest-importance tasks with no observed AI usage in the same data — absence of observed usage, not proof of immunity.
Skills to focus on
Brynjolfsson et al. (2023) found customer service agents using AI saw +14% productivity, with the biggest gains among junior workers — AI compressed the experience gap.
Sources: Felten AIOE (2021), Anthropic Economic Index (2026), Eloundou GPT Exposure (Science, 2024), Pizzinelli et al. bottleneck model. Full methodology.
Singapore Now
Current labour market conditions and how they affect this role.
Cooling, but not collapsing. Vacancies are softer, yet retrenchment remains low and hiring still exceeds resignations.
Vacancy
3.1%
↓ 11.4% YoY
Hiring
2.6%
vs 1.6% resign
Retrenchment
1.5
per 1,000 · low
Re-entry
78.5%
find work in 12mo· -1.6pp
Clerical, Sales & Service Workers · 2025 Q4
Top Industries
Industry vacancy overlays use the latest published detailed cross-tab, which can lag the main labour monitor.
What You Can Do
Healthcare assistant (e.g. therapy aide/assistant) has some offset potential, but it depends on transition pathways holding up in practice and on workers clearing the main switching frictions.
Published transition support
Related roles you could transition to
Similarity-basedCompare within Service & Sales Workers
See how this compares to similar occupations
Compare with... →Classification
Higher risk than 49% of occupations · V7 structural score
Raw scores
AIOE -0.189 · θ 0.674 · C-AIOE -0.153
Stability
watch · Optimistic 10% · Pessimistic 18%
Score range (best/worst case)
Exposure 30–37% · Net risk 11.15–15.96%
Scoring basis
V7 structural score. Uses task-concentration-weighted exposure, human bottleneck, and demand resilience. V6 baseline fields are retained for release-to-release comparison.-0.6pp versus retained V6 baseline.
Wage range (SGD/mo)
25th 2,084 · Median 2,432 · 75th 2,870
Evidence & sources
Data matching
direct · SSOC 53201
Real-world AI usage: +11% vs estimated
Data quality
65% · Matching 0.29 · Market data 0.70 · Freshness 0.80
98% weighted task match · 27% effective coverage
AI overlap by data source
Weights: aioe 32% · anthropic 35% · eloundou 33%
Tools & offset factors
What helps
- Nearby moves and published transition support look reasonably strong.
What could slow it down
- Current demand support is thin, so offsets may take longer to show up.
Worker profile & local context
- Vacancy rate is 3.1% and fell by 0.2 points from last quarter.
- Hiring read: recruitment is running above resignation (2.6% vs 1.6%).
- Retrenchment was low at 1.5 per 1,000 employees.
- 78.5% of retrenched workers re-entered employment within 12 months.
- Live job ads show 0 visible postings in the last 30 days, led by Microsoft Office, Microsoft Excel, Coordination.
Worker profile
Gender mix
15% male / 85% femalePublished Singapore worker composition for the detailed occupation family 53 Personal Care Workers.
Employment structure
Employee-heavy91% employees, 9% employers or self-employed workers.
Work arrangement
Part-time meaningful24% part-time and 76% full-time in 2025.
Age profile
Older-skewing17% aged 15 to 29, 32% aged 30 to 49, and 51% aged 50 or older.
Qualification mix
Non-degree heavySecondary 26%; Post-secondary 25%.
Gross wage by sex
Near parityPublished June 2024 gross wage medians: male $2,386, female $2,450.
Where this work is concentrated
Top planning areas
Woodlands, Tampines, Yishun22% of workers in this occupation group live in these three planning areas.
Residential concentration
Moderately clustered35% live across the top five planning areas in the 2020 Census.
Commute pattern
Mid-range commutesEstimated average commute 33.8 minutes. 28% take 46 minutes or more.
Role profile
How this role's work breaks down across key dimensions. This is a general profile, not an individual measurement.
Workflow dimensions (0 = low, 1 = high)
How this changes by career stage
Senior workers benefit from institutional knowledge and judgment that AI cannot replicate. Entry-level roles have higher task overlap with AI.
Frequently asked questions
Will AI replace Healthcare assistant (e.g. therapy aide/assistant)?
Healthcare assistant (e.g. therapy aide/assistant) has 34% AI task overlap and 50% human bottleneck protection — lower risk than 51% of occupations in the live market. Current AI capabilities have limited overlap with core tasks. Net displacement risk: 14% (Low). Median wage: SGD 2,432/month.
What is the AI risk score for Healthcare assistant (e.g. therapy aide/assistant)?
Healthcare assistant (e.g. therapy aide/assistant) has an AI displacement risk of 14%, rated Low. AI task overlap: 34%. Human advantage: 50%. Local demand buffer: 40%.
What career transitions are available for Healthcare assistant (e.g. therapy aide/assistant)?
Healthcare assistant (e.g. therapy aide/assistant) has modeled transition pathways to related occupations. The strongest adjacent pathway is Nursing aide/assistant, based on skill and wage similarity (model-estimated). Transition scoring accounts for wage preservation, training ease, and destination quality.
How does Healthcare assistant (e.g. therapy aide/assistant) salary compare in the live market?
Healthcare assistant (e.g. therapy aide/assistant) earns a median gross wage of SGD 2,432/month in the live market (25th-75th percentile: SGD 2,084-2,870). This is 46% below median across all 562 scored occupations, and 19% below group median within Service & Sales Workers occupations.