Generative AI changes jobs task by task rather than wiping out whole occupations. Studies by the International Labour Organization, OpenAI researchers and others all start from the same idea: list what a job actually consists of, judge which of those tasks a language model or other AI tool can speed up, and add it up. Writing, coding, data handling and routine paperwork are highly exposed; caring for people, leading teams and physical work in changing environments are not.
This calculator follows that approach with a transparent, simplified model. Each listed role has a typical task mix across six dimensions – digital, analytical, repetitive, human interaction, management and physical work – informed by O*NET task and work-activity descriptions. The first three raise the exposure score, the last three protect it. Small adjustments then reflect your seniority (entry-level work is more routine), your salary relative to the typical US median for the role (higher pay means a stronger business case to automate tasks), how quickly your country and work setting adopt new technology, and whether you work remotely.
The result is an indicator of how much of the job is likely to be reshaped, not a forecast of job loss. Highly exposed jobs usually change by adding AI tools to the workflow; many roles gain productivity rather than disappear.
How to use the AI Job Risk Calculator
- Your role
- Experience level
- Annual salary (before tax) — In US dollars or the equivalent; enter 0 to skip the salary adjustment.
- Country
- Work environment
- Work mode
Everything is calculated in your browser, so results appear instantly and nothing you type is sent to a server.
How it works
Pick the closest role, your experience level, salary (or 0 to skip), country, work setting and work mode. The tool combines the role’s task mix into a base score, applies the adjustments listed in the formula and converts the score into a level and three adoption-speed timelines.
Worked example
An accountant’s task mix in the model is 35% digital, 25% analytical, 20% repetitive, 10% human interaction, 5% management and 5% physical. Exposure E = 35 + 0.8 × 25 + 0.9 × 20 = 73; protection P = 0.9 × 10 + 0.7 × 5 + 5 = 17.5; base score = 100 × 73 ÷ 90.5 = 80.7. A mid-level accountant (±0) earning the median $80,000 (±0) in the United States (+4), in a standard city (±0), working hybrid (+2) scores 86.7 – very high exposure. The expected horizon is 2 + (100 − 86.7) ÷ 100 × 13 = 3.7 years, shortened by 15% for fast-adopting countries to about 3.2 years (1.9 years if adoption is fast, 5.1 if slow).
Score levels and adjustments
| Item | Effect |
|---|---|
| Score below 35 | Low exposure |
| 35–55 | Moderate |
| 55–72 | High |
| 72 and above | Very high |
| Experience | Junior +5 · mid 0 · senior −6 |
| Salary vs median | 6 × log₂(ratio), limited to ±6 |
| Country adoption | High +4 · medium 0 · lower −4 |
| Work setting | Rural −3 · standard 0 · tech hub +3 |
| Work mode | Remote +5 · hybrid +2 · on-site −2 |
Formula
Exposure E = digital + 0.8·analytical + 0.9·repetitive (task shares, %)
Protection P = 0.9·human interaction + 0.7·management + 1.0·physical
Base score = 100 · E ÷ (E + P); Score = base + experience + salary + country + setting + work-mode adjustments (clamped 0–100)
Worked example
Inputs — Your role: Accountant or auditor; Experience level: Mid-level; Annual salary (before tax): $80,000; Country: United States; Work environment: Standard city or town; Work mode: Hybrid.
Results — AI exposure score = 86.7 / 100; Exposure level = Very high; Main exposure driver = Digital work (35% of the task mix); Main protective factor = Human interaction (10% of the task mix); Noticeable task change – fast adoption = around 2028 (≈ 1.9 years); Noticeable task change – expected = around 2029 (≈ 3.2 years).
Key terms
- AI exposure
- The share of a job's tasks that AI tools could perform or substantially speed up.
- Task composition
- How a job's working time divides between digital, analytical, repetitive, interpersonal, managerial and physical activities.
- Augmentation
- AI taking over parts of a job so the worker can focus on other tasks, as opposed to full automation.
- O*NET
- The US Department of Labor database describing the tasks, skills and work activities of about 1,000 occupations.
Frequently asked questions
Will AI take my job?
For most occupations AI automates some tasks rather than the whole job. The ILO's 2025 global index finds about one in four jobs worldwide has some exposure to generative AI, but only a small share – mainly clerical work – is in the highest exposure group.
Which jobs are most exposed to AI?
Jobs made mostly of screen-based, routine information work: data entry, bookkeeping, customer-service chat, administrative support, basic writing and translation. In this model a data entry keyer scores about 90 before adjustments.
Which jobs are safest?
Jobs that combine physical work in unpredictable settings with human contact – plumbers, electricians, nurses, care workers, teachers – score low (about 23–35 base).
Why does a higher salary raise the score?
Expensive tasks give employers a bigger incentive to invest in automation; the adjustment is at most ±6 points (6 × log₂ of salary ÷ median).
Why does remote work raise exposure?
Remote work is by definition digital and well documented, which is exactly the kind of work AI tools handle; fully remote adds 5 points, on-site subtracts 2.
What do the timelines mean?
The years until a noticeable share of the role's tasks is likely to be reshaped by AI tools under fast, expected and slow adoption – not when the job disappears.
How accurate is this?
It is an indicative, simplified model with published weights. Real exposure depends on your specific tasks, employer and regulation; use it to think about which skills to strengthen.