Thematic diagnostic

High exposure often remains augmentation-heavy

The atlas separates how exposed work is from whether AI supports or replaces the worker.

Higher exposure does not automatically imply a replacement-heavy story. The core descriptive object is country-conditioned task exposure. The interpretive question comes next: do exposed tasks lean more toward substitution, augmentation, or a mixed bundle of AI functions?

What changes once we separate the level of exposure from the pathway through which AI enters the task bundle?

Core interpretation

Exposure is not the same object as replacement.

The atlas first measures the share of country-linked tasks that fall into the exposed range. That is already a useful descriptive object, but it is only the first step. A country can rank high on exposure while still leaning toward augmentation rather than direct substitution. The same average can therefore conceal meaningfully different task environments.

What the atlas adds

Pathways tell us how the exposed tasks are exposed.

The pathway layer separates whether exposed work is associated more strongly with substitution, augmentation, or direct execution. That is the layer that turns a country ranking into an interpretable story. Without it, the risk is to read every increase in exposure as if it were evidence of replacement pressure alone.

How to read the evidence

The comparative diagnostics show why averages are not enough.

Kenya and Zambia illustrate that one country can sit higher on the exposure scale while both still remain augmentation-heavy overall. India and Viet Nam show that similar task exposure can coexist with distinct trade-facing integration. The United States and China show that even two high-exposure countries can differ sharply in pathway balance. Together, these comparisons make the same point from different angles: the average level is informative, but the pathway composition is what makes it usable.

Why it matters

Policy questions depend on pathway balance, not just exposure rank.

If exposure is mostly augmentation-heavy, the relevant questions concern complementarity, training, workflow redesign, and organizational adaptation. If substitution-heavy exposure rises, the emphasis may shift toward reallocation, transition support, and distributional effects. The atlas is descriptive rather than predictive, but it helps distinguish those policy conversations more cleanly than a single country ranking can.

Examples

The same rule, seen in different country pairs.

Comparative diagnostic

Kenya and Zambia

Higher average exposure in Kenya does not mean a clean break into a replacement-heavy regime. Both countries still lean more toward augmentation than substitution.

Open the comparison

Comparative diagnostic

India and Viet Nam

The core task benchmark can look similar while the trade-facing layer diverges. Exposure and integration are related, but they are not the same object.

Open the comparison

Comparative diagnostic

United States and China

Both countries sit near the high-exposure frontier, yet China leans more strongly toward substitution. High exposure is not one single pathway regime.

Open the comparison

What to remember

Keep the interpretation disciplined.

The atlas is most useful when the empirical objects stay separate: exposure first, pathway balance second, realized outcomes later.

01

Keep the objects separate

Exposure, pathway balance, and realized adoption are related but distinct objects. The atlas is strongest when those layers are not collapsed into one another.

02

Averages become more useful with pathway detail

Country rankings matter, but they become substantively interpretable only once we inspect which exposed tasks lean toward substitution or augmentation.

03

The evidence is descriptive

These measures show where current AI capabilities map onto country-linked task bundles. They do not, by themselves, identify realized labor-market outcomes or causal effects.

Next steps

Move back into the atlas.

The thematic pages should orient the reader, then route them back to the core atlas and research surfaces.

Read Key Findings Open World View See Methods