← IndicatorsTechnological

Automation Exposure by country

Share of work activities potentially automatable by 2030 (McKinsey Generative AI estimates). Higher = more labour-displacement pressure from AI and automation. Service-heavy economies face higher exposure than agrarian ones.

Global average stress
58.4elevated
Countries covered
25
Unit
%

Global trend

Mean stress score across all countries for which data is available, over the last 12 months.

025456580100Jun 5Jun 6
Bandslowmoderateelevatedhighcritical

Where the pressure is concentrated

The five most-affected and five least-affected countries on this indicator right now.

Most affected

Least affected

Full ranking

All 25 tracked countries, ordered from most to least affected. Click any country to see its full composite breakdown.

#CountryRaw valueStress scoreAs of
1UAE33.0 %88.22023-12-31
2South Korea33.0 %88.22023-12-31
3Singapore32.0 %82.42023-12-31
4Turkey32.0 %82.42023-12-31
5France30.0 %70.62023-12-31
6United Kingdom30.0 %70.62023-12-31
7Israel30.0 %70.62023-12-31
8Poland30.0 %70.62023-12-31
9United States30.0 %70.62023-12-31
10Canada28.0 %58.82023-12-31
11Germany28.0 %58.82023-12-31
12Spain28.0 %58.82023-12-31
13Netherlands28.0 %58.82023-12-31
14Australia27.0 %52.92023-12-31
15Italy27.0 %52.92023-12-31
16Japan27.0 %52.92023-12-31
17Sweden27.0 %52.92023-12-31
18Argentina26.0 %47.12023-12-31
19Mexico26.0 %47.12023-12-31
20Norway26.0 %47.12023-12-31
21Switzerland25.0 %41.22023-12-31
22New Zealand25.0 %41.22023-12-31
23South Africa25.0 %41.22023-12-31
24Brazil24.0 %35.32023-12-31
25India21.0 %17.62023-12-31

Why it matters

Automation Exposure is a contributing indicator to the Technological meta-index, one of the five dimensions of The Human Index composite. Higher raw value = more stress.

Movements in this indicator are tracked daily and feed into every country's composite score on the next cron cycle.

Recent analysis

All Pulses →

Source & methodology

Source: McKinsey Global Institute 2023As of Aging

Raw values are normalized to a 0–100 stress scale per the bounds documented in the codebase. See methodology for the full normalization and band-threshold derivation.