The International Labour Organization (ILO) has released a new research brief examining how artificial intelligence (AI) exposure indicators are used to assess potential impacts on jobs, while also highlighting their limitations. As interest in generative AI grows, these indicators are increasingly used to estimate which tasks and occupations may be automated or transformed, but the ILO cautions that they should not be interpreted as direct predictions of job losses or labour market outcomes.
The brief finds that results vary depending on the methodology used: earlier automation-focused approaches identified lower-skilled, routine jobs as most exposed, while newer AI capability-based measures highlight higher-skilled cognitive roles in sectors such as business, finance, computing, and education. It also notes that exposure is not limited to individual jobs, as highly exposed occupations are often interconnected through shared skills and career pathways, creating broader ripple effects across labour markets.
At the same time, the ILO stresses that all exposure indicators have limitations, as they rely on static task descriptions, do not account for economic feasibility or adoption constraints, and are based on subjective assumptions. They indicate what AI could potentially do rather than what will actually occur. The organization emphasizes that these indicators should be treated as early signals of labour market change and combined with real-world data on employment, wages, and job transitions, along with broader economic and institutional factors, to guide effective and inclusive policy responses.
