New OECD measure compares AI and job capabilities
Office and administrative support occupations show the closest match with current AI capabilities, according to the OECD.
The OECD has published a new framework designed to assess how closely current AI capabilities align with the requirements of different occupations.
The paper, ‘The OECD AI Exposure Measure‘, maps OECD AI Capability Indicators to occupations and introduces an AI Capability Gap Index. According to the OECD, the framework is intended to support analysis of potential AI impacts on work, skills, education, and labour-market policy.
The framework compares AI capabilities with occupational requirements across nine domains: language, social interaction, problem-solving, creativity, metacognition and critical thinking, knowledge, learning and memory, vision, manipulation, and robotic intelligence. Occupations with smaller capability gaps are considered more exposed to current AI capabilities, while larger gaps indicate a greater distance between AI systems and occupational requirements.
The OECD emphasised that the measure is not intended as a prediction of automation or job loss. It measures potential exposure to current AI capabilities, while actual labour-market effects will also depend on adoption, costs, task structure, regulation, organisational uptake, and social choices.
The report found that occupations involving routine information processing and administrative tasks currently show the highest levels of AI exposure. Office and administrative support occupations record the lowest total gap index, followed by production, food preparation and serving, and sales-related occupations.
Occupations relying more heavily on judgement, social interaction, interpretation, and non-standardised physical activity showed larger capability gaps.
The paper also noted that different forms of AI may affect occupations differently depending on whether work relies more on reasoning, communication, robotics, or physical interaction.
The OECD said the framework could support future task-level analysis, scenario modelling, and country-specific assessments of AI-related labour-market change. Future work may extend the approach to task-level analysis, scenario applications, macroeconomic modelling, and country-level assessments.
Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!