IoM Workforce — Australian AI Risk Model
Isle of Man Census workers grouped by SOC 2020 occupation category, cross-referenced against the Jobs and Skills Australia Gen AI Capacity Study. Each IoM occupation group is mapped to the equivalent Australian ANZSCO group (both systems derive from ISCO-08), then Australian automation and augmentation scores are applied to the IoM headcount to estimate workforce exposure.
Australian JSA automation scores applied to IoM Census headcounts. Cell size = number of IoM workers. Red = most tasks at automation risk.
Fraction of tasks in each group that Gen AI could perform autonomously (Australian JSA model). Red = most tasks at risk of automation.
Methodology & Crosswalk
Why Australian data? The JSA Generative AI Capacity Study is one of the most comprehensive task-level AI exposure analyses available, covering 358 ANZSCO occupations with scores derived from the ILO Gen AI methodology. It provides a rich, research-backed proxy for occupational AI risk.
SOC 2020 → ANZSCO mapping: Both the UK Standard Occupational Classification (SOC 2020) and the Australian and New Zealand Standard Classification of Occupations (ANZSCO v1.3) are derived from the International Standard Classification of Occupations (ISCO-08). At the major group (1-digit) level, the correspondence is well-established in published ABS, ONS and ILO crosswalk tables. Note: SOC groups 3 and 5 (Associate Professional & Technical, and Skilled Trades) both map to ANZSCO group 3 (Technicians and Trades Workers).
Score computation: Australian scores are weighted averages across all ANZSCO unit groups in each major group, weighted by the number of tasks assessed per occupation. These are then applied to the IoM Census respondent count for the equivalent SOC major group.
Limitations: This is a cross-national proxy. The IoM labour market may differ from Australia in occupation mix, skill levels, and task composition within each major group. Scores should be interpreted as directional indicators, not precise IoM-specific measurements.
SOC 2020 → ANZSCO Major Group Crosswalk
| SOC | UK SOC 2020 Group | → ANZSCO | Australian ANZSCO v1.3 Group |
|---|---|---|---|
| 1 | Managers, Directors and Senior Officials | 1 | Managers |
| 2 | Professional Occupations | 2 | Professionals |
| 3 | Associate Professional and Technical Occupations | 3 | Technicians and Trades Workers |
| 4 | Administrative and Secretarial Occupations | 5 | Clerical and Administrative Workers |
| 5 | Skilled Trades Occupations | 3 | Technicians and Trades Workers |
| 6 | Caring, Leisure and Other Service Occupations | 4 | Community and Personal Service Workers |
| 7 | Sales and Customer Service Occupations | 6 | Sales Workers |
| 8 | Process, Plant and Machine Operatives | 7 | Machinery Operators and Drivers |
| 9 | Elementary Occupations | 8 | Labourers |
* SOC 3 and SOC 5 both map to ANZSCO 3 — the UK separates technical associate professionals from skilled trades, while ANZSCO combines them.
IoM Census data: dbo.vw_census_treemap. Australian exposure data: dbo.jsa_ai_exposure (import via pnpm scrape:import-jsa).
Compare with IoM-only analysis at /census (Anthropic Economic Index + Felten AIOE data).