Malaysia spared from generative AI job losses so far
Middle- and low-income countries have seen little change in labour demand after the launch of ChatGPT, the ASEAN+3 Macroeconomic Research Office reported on Wednesday.
Source: ASEAN+3 Macroeconomic Research Office · July 23, 2026 at 11:07 PM · AI-assisted report
Single-source
KUALA LUMPUR, 24 JULY 2026 —
Middle- and low-income countries have seen little change in labour demand after the launch of ChatGPT, the ASEAN+3 Macroeconomic Research Office reported on Wednesday.
Market Impact
The AMRO Café seminar paper, using Lightcast data for 555 million job postings across 84 countries from 2021 to 2025, found postings for high-vulnerability occupations in high-income economies fell 5.8% relative to low-vulnerability roles after November 2022. In middle- and low-income economies the effect was statistically insignificant, AMRO said.
The authors, Dr. Shu Yu of the World Bank and Dr. Xianguo (Jerry) Huang of AMRO, attribute the difference to lower AI adoption, weaker complementary skills and less exposure to digital services trade in lower-income countries. They use a difference-in-differences design that controls for GenAI exposure and monetary policy sensitivity.
For Malaysia, the findings imply limited near-term displacement risk because the country ranks below high-income peers in GenAI uptake and digital services specialisation. Bank Negara Malaysia’s 2024 annual report shows AI-related investment accounted for less than 0.5% of gross fixed capital formation last year.
The study covers postings data only through 2025Q2 and does not include Malaysia-specific sub-samples.