ThisCalculator

Everyday

AI Automation Risk: How to Assess Any Job

Updated · Standard formula, shown below

This guide explains a practical method for evaluating how exposed any job is to AI automation. You’ll learn to break a role into its core tasks, score each task against known automation risk factors, and combine those scores into an overall exposure rating. No special tools are needed—just a clear head and a basic spreadsheet or pen and paper.

What You’ll Need

  • A list of the job’s main tasks (from a job description, your own notes, or a colleague’s input)
  • A spreadsheet program (Excel, Google Sheets) or just paper and a pen
  • About 30 minutes per job role

Step-by-Step

  1. List the core tasks. Write down 8–15 specific tasks that make up the job. Be concrete: instead of “manage projects,” write “create project timelines and update status reports.”
  2. Rate each task for routine-ness (1–5). A score of 1 means the task is highly variable, creative, or requires complex judgment (e.g., “negotiate a contract with a new client”). A 5 means the task is repetitive and predictable (e.g., “enter invoice data into the accounting system”).
  3. Rate each task for data reliance (1–5). How much does the task depend on processing structured data (numbers, text, images)? 1 = almost no data, 5 = almost entirely data-driven (e.g., “classify customer emails by topic”).
  4. Rate each task for physical interaction (1–5). 1 = no physical presence or manipulation needed (e.g., “write code”), 5 = requires hands-on work in an unpredictable physical environment (e.g., “repair a washing machine”).
  5. Rate each task for social/emotional requirement (1–5). 1 = no human interaction or empathy needed, 5 = requires deep empathy, persuasion, or trust-building (e.g., “counsel a grieving family member”).
  6. Calculate a task automation risk score. For each task, use this formula: Risk = (Routine + Data) – (Physical + Social). The result will range from -8 to +8. A positive score means the task is more exposed to automation; a negative score means it’s safer.
  7. Weight by time spent. Estimate what percentage of the workweek each task takes. Multiply each task’s risk score by that percentage (as a decimal). Sum all the weighted scores to get the job’s overall automation exposure index.
  8. Interpret the index. A score above +3 suggests high exposure; between -3 and +3 is moderate; below -3 is low. But don’t treat these as absolute thresholds—use them as a relative ranking when comparing different jobs.

Common Mistakes

  • Treating the job title as the unit of analysis. “Accountant” can range from data-entry clerk to strategic CFO. Always assess the actual tasks, not the label.
  • Ignoring the context. A task that seems routine (e.g., “drive a delivery truck”) may involve unpredictable traffic, weather, and customer interactions that lower its automation risk.
  • Assuming AI does the whole job. Many jobs are a mix of automatable and non-automatable tasks. Automation often replaces tasks, not entire roles.
  • Forgetting about cost and adoption. A task may be technically automatable, but if the technology is expensive or the company lacks the skills, it won’t happen soon. Your score measures potential, not timeline.

Tips

  • Use a 1–5 scale consistently. If you’re unsure between two numbers, pick the higher risk (more routine, more data) to be conservative.
  • Get a second opinion. Ask someone who does the job to review your task list and scores—they’ll catch things you missed.
  • Focus on the highest-weighted tasks. A single task that takes 60% of the time can dominate the overall score, so double-check those percentages.
  • Compare jobs within the same industry. An index of 2 might be high for a nursing role but low for a data-entry role.

Troubleshooting

If your scores all cluster near zero: You may be rating tasks too broadly. Break them down further. For example, instead of “handle customer inquiries,” split into “answer FAQs” (high routine) and “resolve unique complaints” (low routine).

If a job scores high but you think it’s safe: Check if you underestimated the physical or social components. Many jobs have hidden human elements—like a janitor who notices when a hallway is unsafe—that aren’t in the official task list.

If you’re comparing many jobs and they all look similar: Add a fifth factor: “adaptability.” If a job requires learning new skills frequently, that reduces its exposure because AI systems are usually trained on fixed patterns. Score adaptability 1 (low) to 5 (high) and subtract it from the risk score.

Frequently Asked Questions

Does a high score mean the job will disappear?

No. It means the tasks are likely to be automated, but the job may evolve to focus on other tasks. For example, bank tellers lost routine transactions but gained advisory roles.

Can I use this for my own career planning?

Yes. If your job scores high, look for ways to shift your time toward tasks with lower automation risk—like complex problem-solving or relationship management.

Is this method scientifically proven?

It’s based on widely cited research (like Frey & Osborne’s 2013 study) that identifies routine, data-heavy tasks as most automatable. But it’s a simplification; real-world outcomes depend on many factors.

What about AI that can write code or create art?

Those tasks may score high on routine/data if they’re well-defined (e.g., “generate a standard contract”). But creative work that requires original thought and taste still scores low on routine.

How often should I reassess?

Every 6–12 months, because AI capabilities change quickly. Re-run the assessment with updated task descriptions and note any new tools that have become common in the industry.

Can I use this to compare two different job offers?

Yes. Run the assessment on each role and compare the overall indices. The one with a lower index is likely more resilient to automation in the long term.

Related

Method: the standard formula for this calculation, rounded as each result states. See how the calculators work or report an error.