Macro Markets and Machines_The Economic and Market Transformation Driven by AI_GWM report - Flipbook - Page 39
Macro, Markets, and Machines
November 2025
Section 4.8: Employment Effects Will Depend on Complementarity with AI
The IMF and BIS do not detail the impact of greater AI adoption on aggregate employment given simplified
assumptions on labour markets where workers reallocate or the labour force adapts over the medium term to
the demands of higher AI adoption. In the near term, the current composition of labour markets and the skills
and adaptability of workers could result in job losses for some, with certain demographic groups possibly
struggling with new technology and facing weaker re-employment prospects following termination.
Demographic implications may also be more broad based than in previous waves of technological innovation,
such as the Industrial Revolution, which had a greater effect on the male workforce and in certain regions more
than others.
Within organizations and across industries, some roles or functions have a higher degree of exposure to AI
split across its augmentative power or its automation risks (Exhibit 33). The net impact on employment would
also have important implications for aggregate demand, which may act as a disinflationary or inflationary force
that complements the BIS’s and IMF’s results in estimating the central bank response.
Exhibit 33 – Job Function Exposure to Automation and Augmentation from LLM
100%
90%
26
21
80%
16
70%
60%
50%
28
41
34
41
18
41
50
21
18
40%
35
34
27
30%
20%
14
44
42
33
32
10%
29
22
22
19
HR
Marketing
Legal
0%
IT/Tech
Finance
Automation
Cust. sales
Augmentation
Operations
Lower potential
Supply chain
Non-language tasks
Sources: World Economic Forum; Scotia Wealth Management.
Cross-country differences in exposure and potential complementarity with AI would determine the balance
of risks and benefits to their labour from AI. In advanced economies, a higher proportion of jobs have
significant exposure to AI, which would translate into bigger aggregate productivity benefits, but this higher
exposure may also increase risks of displacement by AI depending on whether the technology complements or
replaces workers, with the opposite being true for lower-income countries where AI-intensive or beneficiary
sectors represent a smaller share of the labour force or economy. A higher share of AI exposure cuts both ways,
with advanced economies having more workers at each side of the positive- and negative-complementarity
spectrum, with respective gains and losses from the adoption of AI that point to a more polarized impact in
within-country labour markets than low-income countries, with implications for income inequality and/or public
policy. As Exhibit 34 shows, U.S. workers have a higher familiarity with AI skills, which may allow them to
capitalize on the complementarity benefits of the technology with greater ease than those in other countries
where a knowledge gap could favour automation.
Scotia Wealth Management
38
