To tell which work is most exposed to AI automation, assess the tasks in a job against a clearly defined AI capability—not the job title alone. Then state how task-level findings are combined, distinguish technical exposure from likely adoption, and check real employment evidence before drawing conclusions about job loss. An exposure measure describes what AI might do, not what will happen to a worker.
What “AI exposure” measures—and what it does not
Exposure is a relationship between a set of work tasks and a specified technology. A language model that drafts text, software that analyzes information, and a robot that handles physical objects have different capabilities and affect different tasks. A result is meaningful only when it makes clear which technology it covers and whether it assesses current tools or a forward-looking scenario.
Exposure is not a probability that a person will be laid off, an estimate of unemployment, or a ranking of jobs that are safe or doomed. Even if a system can perform or accelerate a task, employers may not adopt it: costs, workflow redesign, regulation, accountability, and demand all matter. The International Labour Organization (ILO) cautions that indicators capture what AI could do, “as a first step in the analysis, not what will happen in practice.” ILO, 17 April 2026.
For personal-finance decisions, that distinction matters. An exposure score alone cannot tell you whether your income is at risk, whether your employer will change staffing, or whether a role’s demand will grow or shrink. Those questions require evidence about adoption and labor-market outcomes, not only technical capability.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
#1 Best Overall
How to assess a job’s exposure
- Set the scope. Specify the geography, occupation classification, technology, and time horizon. For example, a U.S. estimate based on O*NET task descriptions is not automatically a global result. The ILO’s global occupational analysis maps work using ISCO-08; OECD studies may use other data and methods.
- Break the job into actual tasks. Include routine information processing, communication, analysis, judgment, care, physical handling, and accountability where they apply. One job title can cover very different task mixes across employers and workers.
- Test each task against the defined capability. Ask whether the system can complete the task, speed it up, or assist only one step. Note the degree of human review assumed. Some methods count exposure when an LLM could complete a task in half the time; that threshold is a methodological choice, not a universal definition.
- Explain the aggregation. Say how task results become an occupation-level measure—for instance, the share of tasks above a threshold, a mean exposure score, or a capability gap. Preserve differences within an occupation when the measure provides them instead of presenting an average as though every worker does the same work.
- Assess adoption separately. Consider whether use is economically and organizationally feasible, and whether regulation, responsibility, or institutional barriers constrain it. Technical capability does not establish actual workplace use.
- Check outcomes before claiming job losses. For claims about employment effects, examine observed employment, wages, hiring, and worker transitions. Do not substitute an exposure index for those measures.
Why rankings differ across studies
Two studies can produce different rankings without either being mistaken: they may define exposure differently, assume different AI capabilities, use different task data, or aggregate results in different ways. Compare the methods before comparing headline percentages.
| What to compare | Why it matters |
|---|---|
| Technology included | An LLM or generative AI measure, broader AI capabilities, physical robotics, and combined systems do not describe the same task set. |
| Definition of exposure | Studies may measure task completion, time saved, capability overlap, or another operational definition. A threshold can change which tasks count. |
| Time horizon | Current capability and plausible near-future tools answer different questions. |
| Task and occupation data | Country, occupational classification, task-description source, granularity, and date affect what is represented. O*NET and ISCO-08 are not interchangeable taxonomies. |
| Aggregation | A task-share threshold, mean score, or capability gap can yield different occupation results. Check whether within-occupation task variability is represented. |
| Outcomes measured | A technical exposure measure does not show employment, wage, or transition effects unless the study separately measures those outcomes. |
The ILO’s 2026 brief notes that indices can rely on static task descriptions, subjective assumptions, and omit constraints on adoption. It also warns that results vary with the measure used. Avoid treating percentages from different methods as if they were directly comparable.
Rank #2
- Ideal for Gifting
- Ideal for a bookworm
- Compact for travelling
What recent evidence says about exposure
The ILO’s 2025 refined global index estimates that one in four workers worldwide are in an occupation with some degree of generative AI exposure. That is an exposure estimate, not a prediction that one in four workers will lose a job. In the same index, 3.3% of global employment falls into its highest exposure gradient; the reported shares are 4.7% of female employment and 2.4% of male employment, with differences varying by country income. These figures depend on the ILO’s method and occupational mapping, not a universal ranking of individual jobs. ILO, “Generative AI and jobs: A 2025 update”; ILO working paper, 20 May 2025.
The ILO’s updated index also changed the distribution of its scores: its reported mean automation score was 0.29 in 2025 versus 0.30 in 2023, while the standard deviation was 0.14 versus 0.30. These are changes in the index under an updated methodology, not observed changes in employment. The ILO’s paper combines mean exposure with task variability in four gradients, rather than treating all jobs with a similar average as identical.
Recommended Free Tools
Exposure is not confined to work commonly described as low-skill. Clerical occupations score highly in the ILO analysis, while newer capability-based measures also identify exposure in professional and cognitive fields such as business, finance, computing, and education. Which occupations appear most exposed depends on the technology and measurement method. Earlier rankings focused on routine-task automation should not be assumed to apply unchanged to generative AI.
The OECD’s 2026 AI exposure measure maps capabilities across nine cognitive, social, and physical domains to occupational requirements. It is a forward-looking capability measure, not an estimate of observed job losses. OECD, “The OECD AI exposure measure” (2026). A separate OECD framework examining generative AI uses task-share approaches and distinguishes exposure now from exposure now or in the near future. OECD, “Beyond automation” (2024).
Rank #4
When to include robotics
If the question is specifically about generative AI or language models, keep the analysis to those systems. If it is about AI automation broadly, say whether physical robotics is included. Robots can affect physical tasks that a text-focused exposure measure does not capture, so combining their effects without defining the scope can obscure rather than clarify the result. Anthropic’s 2026 work on robot exposure illustrates a separate approach focused on physical tasks. Anthropic, “Can we predict the jobs robots will do?” (2026).
How to use exposure information for your own work
Use published occupational measures as a prompt to examine your own task mix, not as a personal forecast. List the recurring tasks in your role and note which involve producing or processing information, interacting with people, exercising judgment, handling physical objects, or carrying responsibility for outcomes. Then ask whether the relevant AI can complete the task, accelerate part of it, or only support a human doing it.
Free tools Windows power users keep installed
One-click scans. No signup required.
Quick Recap
Best Value
- It can be a gift option
- Comes with secure packaging
- Helpful in various ways
- Look for changes in how your employer assigns work, hires, trains, or evaluates performance; those are adoption signals, not proof that a job will disappear.
- Separate automatable steps from the broader service or outcome your role provides. A task may change while the job is reorganized around review, exceptions, coordination, or accountability.
- For financial planning, avoid making a large decision—such as assuming a role is secure or abandoning a career path—based on one exposure score. Pair it with local hiring, wage, and transition evidence relevant to your occupation and region.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
