task exposure
In plain wordsA measure of how much of a job's work AI could, in principle, do or speed up. High exposure means many of the job's tasks are within AI's reach, not that the job will vanish.
An estimate of the degree to which the tasks making up an occupation could be performed, or substantially sped up, by AI. Studies typically break occupations into task statements from a taxonomy such as the US O*NET database, rate each task's susceptibility using expert or model judgments, or overlaps between AI capabilities (or patents) and task descriptions, and aggregate by occupation, weighting tasks by importance or time. A widely cited 2023 study estimated that large language modelslarge language model, n.: A computer program trained on enormous amounts of text to predict the words that come next. At vast scale, that simple skill lets it chat, answer questions, translate, summarize and write code. could reach at least a tenth of the tasks of about 80% of US workers, and at least half the tasks of about 19%. Exposure measures technical potential, not outcomes: it does not say whether exposure will mean automationautomation, n.: Using machines or software to do work that people would otherwise do, from factory robots to programs that sort email or write reports. or augmentationaugmentation, n.: Using AI to help people do their work better or faster while they stay in charge, instead of having the AI do the job on its own., how fast adoption will be, or how wages and employment will respond. Usage-based measures drawn from real assistant conversations complement it.
Translators score high on task exposure, but so do many higher-paid financial analysts.
Origin
Builds on the task-based approach of David Autor, Frank Levy and Richard Murnane (2003). AI-specific measures followed from Erik Brynjolfsson, Tom Mitchell and Daniel Rock (suitability for machine learning, 2018), Edward Felten, Manav Raj and Robert Seamans (AI occupational exposure, 2018–2021) and Michael Webb (2019); Tyna Eloundou and colleagues' 'GPTs are GPTs' (2023) made exposure central for language models.
Usage note
Scores differ markedly across methods, so rankings of the most exposed jobs are rough. High exposure has been associated both with productivity gains and with reduced hiring of entry-level workers.
Related
- technological unemploymenttechnological unemployment, n.: Joblessness caused by machines or software taking over work that people used to be paid for, faster than new kinds of work appear for them.
- automationautomation, n.: Using machines or software to do work that people would otherwise do, from factory robots to programs that sort email or write reports.
- augmentationaugmentation, n.: Using AI to help people do their work better or faster while they stay in charge, instead of having the AI do the job on its own.
- jagged frontierjagged frontier, n.: The uneven edge of what AI can do: it may excel at some hard tasks yet fail at seemingly easier ones, so how hard a task looks is a poor guide to whether AI can handle it.
- general-purpose technologygeneral-purpose technology, n.: A technology, like electricity or the computer, that spreads through the whole economy, keeps improving and sparks many other inventions. Many economists think AI is one.
- GDPvalGDPval, prop. n.: An OpenAI test that pits AI against seasoned professionals on genuine workplace assignments, from reports and spreadsheets to slide decks, in 44 occupations important to the US economy.
Sources
- Eloundou, T., Manning, S., Mishkin, P. & Rock, D. (2024). GPTs are GPTs: Labor market impact potential of LLMs. Science 384(6702), 1306–1308.
- Felten, E., Raj, M. & Seamans, R. (2021). Occupational, industry, and geographic exposure to artificial intelligence: A novel dataset and its potential uses. Strategic Management Journal 42(12), 2195–2217.
Mentioned in
automationautomation, n.: Using machines or software to do work that people would otherwise do, from factory robots to programs that sort email or write reports., GDPvalGDPval, prop. n.: An OpenAI test that pits AI against seasoned professionals on genuine workplace assignments, from reports and spreadsheets to slide decks, in 44 occupations important to the US economy., general-purpose technologygeneral-purpose technology, n.: A technology, like electricity or the computer, that spreads through the whole economy, keeps improving and sparks many other inventions. Many economists think AI is one., productivity paradoxproductivity paradox, n.: The puzzle that a powerful new technology, like computers in the 1980s or AI today, can be everywhere without showing up as faster growth in how much workers produce., technological unemploymenttechnological unemployment, n.: Joblessness caused by machines or software taking over work that people used to be paid for, faster than new kinds of work appear for them., Erik BrynjolfssonErik Brynjolfsson, prop. n.: An American economist who studies how digital technology and AI change productivity and jobs, and who warns against AI built to replace people rather than to augment them., Ethan MollickEthan Mollick, prop. n.: An American management professor at Wharton whose experiments and popular writing on how people use AI at work introduced ideas such as the 'jagged frontier'., Remote Labor IndexRemote Labor Index, prop. n.: A test that gives AI agents real paid freelance projects — design, video, data work and more — and checks how many they finish to a standard a client would accept. Early agents managed very few., Tom MitchellTom Mitchell, prop. n.: An American computer scientist who wrote a classic machine-learning textbook with the field's most quoted definition of learning, and founded the first university department devoted to machine learning.