Benchmarking the Future of Work: Mapping AI Progress to Occupational Tasks

Abstract

Artificial intelligence is advancing rapidly, yet we still lack tools for mapping these gains onto the world of work, and in particular onto occupations’ exposure to AI. Existing measures of AI exposure rely primarily on expert judgments, patent text, static task classifications, or observed patterns of use. This paper introduces the Benchmark-Anchored Index of Occupational Exposure (BAIOE), a new methodological framework that maps frontier technological progress to occupations by using O*NET abilities as a bridge. The approach links occupations to benchmarks - the scoreboards that track frontier AI capabilities - and thereby produces a measure of exposure that is grounded in demonstrated technological performance, dynamically updatable as new benchmarks emerge, and forward-looking in its ability to identify where automation pressures are likely to arise. By repositioning benchmarks from technical scoreboards to economic indicators, this study offers a fresh lens for anticipating the future of work and shaping policy responses.

Keywords: innovation