How Much Do You Know About AI and Jobs?
Eight questions on how AI exposure is measured.
Take the quizType your job title to see how much of your work today's AI can speed up, task by task, based on U.S. Department of Labor data and published research.
Loading ~900 occupations… Use the arrow keys to move through suggestions and Enter to pick one.
Data as of Sep 2026 · Sources: O*NET task statements as used in Eloundou et al. (2023), “GPTs are GPTs” (2023 ratings), O*NET 31.0 skills and related jobs (Aug 2026), Anthropic Economic Index (Mar 2026)
Shares are weighted like the research: core tasks count double.
Uncheck anything you don’t actually do. Your score updates to match your real job.
Picked from the 35 O*NET skills whose importance doesn’t rise with AI exposure. Ratings run from 1 (not important) to 5 (extremely important).
Drawn from O*NET’s related-occupation lists. Tap one to check it.
Eyeing a lower-exposure move? Compare salary, training costs and lifetime earnings.
Weighing an offer in a safer role? Compare total comp side by side.
If a layoff comes, know what a fair severance package looks like.
Staying and upskilling? Build a defensible number for your next raise.
1. Tasks. Each occupation’s task list comes from the O*NET task statements as used in Eloundou et al. (2023), “GPTs are GPTs”: 923 occupations and about 19,000 task statements from the U.S. Department of Labor’s O*NET. These are the task lists the researchers rated, so a few can differ slightly from the newest O*NET release.
2. Exposure. Every task was rated in “GPTs are GPTs” (Eloundou, Manning, Mishkin & Rock; arXiv 2023, Science 2024) by trained human annotators and, separately, by GPT-4. The question was whether a large language model could cut the time the task takes by at least half at the same quality. The ratings are not exposed (0), exposed with extra LLM-powered software (0.5) or directly exposed (1). Your score averages the human and GPT-4 ratings across your tasks, with core tasks counted double (the paper’s β measure), then scales it to 0–100.
3. Buckets. Automatable now means both raters judged the task exposed, at least one said an LLM alone is enough, and either GPT-4 rated it for significant or complete automation or Anthropic observes it in real AI usage. AI-assisted means AI can speed the task up, but a person stays in the loop. Human-led means neither rater found meaningful exposure (or only one did, and only with extra tools). If a task shows up in Anthropic’s real-world Claude usage data and at least one rater found some exposure, we label it at least AI-assisted, even if the research ratings alone would call it Human-led.
4. Real-world use. The “Seen in real AI use” task flags and the occupation-level “Seen in real AI use today” figure (Anthropic’s “observed exposure”) come from the Anthropic Economic Index Labor market impacts data (Massenkoff & McCrory, 2026), which is based on how Claude is actually being used at work.
5. Skills and moves. Suggested jobs are O*NET related occupations that score at least 8 points lower, and usually at the same or a higher Job Zone. Skills are the O*NET skills those jobs need more than yours does, limited to skills whose importance doesn’t rise with AI exposure among jobs at the same Job Zone.
Limits. Exposure measures what AI could speed up, not whether your employer will cut jobs. The ratings reflect early-2023 AI capabilities, so newer models and agents probably push many scores higher. Task lists describe the typical U.S. worker, not your specific role. Treat the result as a conversation starter, not a forecast.
Sources & licenses: Task lists: O*NET task statements as used in Eloundou et al. (2023), “GPTs are GPTs” (arXiv 2303.10130). · Occupation descriptions, job titles, Job Zones, skills and related occupations: O*NET 31.0 Database by the U.S. Department of Labor, Employment and Training Administration (USDOL/ETA), used under CC BY 4.0. It was modified here: tasks were matched to exposure ratings and scores were computed. USDOL/ETA has not approved, endorsed or tested these modifications. O*NET® is a trademark of USDOL/ETA. · Eloundou, T., Manning, S., Mishkin, P., & Rock, D. “GPTs are GPTs”, Science 384 (2024); data from github.com/openai/GPTs-are-GPTs (MIT License, © 2024 OpenAI). · Anthropic Economic Index, labor_market_impacts (Massenkoff & McCrory, “Labor market impacts of AI”, 2026), data released under CC-BY.
Companion pieces
Quizzes, guides and answers built to go with this calculator.
Eight questions on how AI exposure is measured.
Take the quizProbably not in one step. For most occupations, AI can speed up some tasks while others still need a person’s judgment, hands, presence or accountability. A high score means your job is likely to change a lot. Routine output shifts to AI, and the value moves to review, relationships and the human-led tasks. Anthropic’s 2026 labor-market study found no systematic rise in unemployment in highly exposed jobs so far. It did find early signs of slower hiring of 22–25-year-olds into those jobs.
It’s the weighted share of your tasks that a large language model could do at least twice as fast. A directly exposed task counts fully, a task that needs extra AI-powered software counts half, and core tasks count double. It comes from the “GPTs are GPTs” ratings (Eloundou et al.). 0 means no exposed tasks. 100 means every task is directly exposed.
Task lists are the O*NET task statements as used in Eloundou et al. (2023), “GPTs are GPTs”. Task exposure ratings come from that paper’s open dataset (MIT License). Job titles, skills and related occupations come from the U.S. Department of Labor’s O*NET 31.0 database (CC BY 4.0). Real-world usage signals come from the Anthropic Economic Index (CC-BY). Everything is precomputed. Nothing you type is sent to an AI model or stored.
Hands-on, in-person work scores lowest: skilled trades such as electricians and plumbers, equipment installation and repair, construction, food preparation, and many direct patient-care roles. Across all occupations, skills like operating, maintaining, repairing and troubleshooting equipment are the most AI-resilient. Coordinating people and reading social cues also hold up well.
Lean into the tasks marked human-led, use AI to take over the automatable ones rather than competing with it, and build the listed skills. If you want to switch roles, the “similar jobs” list shows related occupations with lower exposure that use much of what you already know.
It’s an estimate of how exposed a job’s tasks are, not a prediction about your employer. The ratings reflect early-2023 AI capabilities, so exposure has likely grown since. Your own mix of tasks matters, so uncheck the tasks you don’t do to personalize the score.
8 questions, about 4 minutes, instant results and no sign-up.