ILO warns against treating AI exposure indicators as job-loss forecasts

A new brief from the International Labour Organisation argues that AI exposure indicators should not be treated as forecasts of job losses, even as they become a more common tool for assessing how artificial intelligence could reshape work.

According to the ILO, these indicators can help identify where jobs may be affected by AI. Still, they do not show whether workers will actually be displaced or how labour markets will adjust in practice.

The brief examines how different exposure measures are constructed and why they often produce different results. Earlier approaches to automation focused mainly on routine and lower-skilled work, while newer AI-related models point to greater exposure in higher-skilled cognitive occupations, including roles in finance, computing, business, and education. That shift reflects the growing capacity of AI systems to perform tasks once seen as less vulnerable to automation.

The ILO stresses that exposure does not necessarily lead to job loss. Most indicators rely on static task descriptions and estimate what may be technically feasible, rather than what employers will actually adopt or what makes economic sense. They do not capture whether automation is profitable, whether it improves productivity, or how firms, workers, and institutions may respond over time.

The brief also argues that AI-related disruption is unlikely to stay confined to a narrow set of occupations. Jobs are linked through shared skills, career mobility, and workplace structures, meaning that changes in one part of the labour market can influence broader employment patterns elsewhere. That makes simple occupation-by-occupation risk scores less useful on their own than they may appear.

For that reason, the ILO says exposure indicators should be used as early warning signals rather than stand-alone labour market forecasts. It recommends combining them with evidence on employment, wages, job transitions, and broader economic and institutional conditions to build a more realistic picture of how AI is affecting work.

The broader significance of the brief is that it pushes back against the simplest narratives about AI and employment. Rather than asking how many jobs AI will eliminate, the ILO is urging policymakers to focus on where work may change, how quickly adoption may happen, and what kinds of institutions, skills, and labour protections will shape the outcome.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!