Calculators

AI and automation

AI Job-Risk Score Calculator

Estimate whether your job tasks are exposed to AI, which parts are most vulnerable, and what to learn next.

When to use this

Use it when the question is really about tasks, not titles

Two people can share the same job title and have very different AI exposure. This tool scores the work mix you enter and shows which task groups are easiest to redesign with AI.

Default result

The server-rendered Marketing Operations Specialist example scores 53.8%: gradient 3: significant exposure.

Score your task mix

Use the presets to choose an exposure score, then set whether each task is core or supplemental and how much of the role it represents.

Task 1
Task 2
Task 3
Task 4
Task 5
Task 6
Score preset meanings
  • 0.90: Direct AI fit: writing, translation, summarizing, code, data cleanup
  • 0.72: High AI assist: reporting, dashboards, research synthesis
  • 0.55: Tooling-enabled: routing, follow-up, support drafts, CRM work
  • 0.34: Mixed: ambiguous escalation, judgment-heavy coordination
  • 0.22: Lower fit: negotiation, live trust, sensitive stakeholder work
  • 0.18: Lowest fit: final strategy calls, physical or in-person accountability

Uncertainty

Sigma 0.224

Higher sigma means the role mixes high-exposure and low-exposure tasks.

Risk pattern

Dashboard and summary work lead exposure

The highest contribution comes from high-score tasks with core weighting.

What to learn next

Use AI as a support layer while strengthening trust, domain judgment, stakeholder work, and accountability-heavy parts of the role.

Worked example: Marketing Operations Specialist AI exposure

The default role has three core tasks and three supplemental tasks. Core tasks carry weight 2, supplemental tasks carry weight 1, so total weight is 2 + 2 + 2 + 1 + 1 + 1 = 9.

The weighted exposed score is 2 x 0.78 + 2 x 0.72 + 2 x 0.55 + 1 x 0.34 + 1 x 0.22 + 1 x 0.18 = 4.84.

Exposure is 4.84 / 9 = 0.5378, displayed as 53.8%. The weighted standard deviation is 0.224.

Because exposure is between 0.5 and 0.6, and exposure plus sigma is above 0.5, the default result is Gradient 3: significant exposure.

At-risk share is the task weight with scores of 0.67 or higher: the first two core tasks contribute 4 / 9 = 44.4%. Transform share is 2 / 9 = 22.2%. Safer share is 2 / 9 = 22.2%.

How the AI job-risk score is calculated

The score is a weighted task exposure estimate. Each task has an AI exposure score from 0 to 1, a task type weight, and a work-share multiplier. Core tasks count twice as much as supplemental tasks.

The formula is exposure equals sum of weight x score divided by sum of weight. Sigma measures how spread out the task scores are around the weighted average.

What makes a task high exposure

Writing, summarizing, translation, code, spreadsheet work, document analysis, and repeatable software-mediated tasks usually score higher because current AI tools can draft, transform, classify, or route the work.

Why this is directional, not a forecast

The score measures technical task exposure. It does not predict layoffs, wages, promotion odds, employer adoption, legal limits, data access, or whether a human remains accountable for the final work.

At-risk tasks vs safer tasks

At-risk share counts task weight with scores of 0.67 or higher. Transform share counts scores from 0.50 to below 0.67. Safer share counts scores below 0.33.

What to learn next, by task pattern

If writing, analysis, or admin tasks dominate the score, learn AI workflow design, prompt review, source verification, QA, and exception handling. If relationship or judgment tasks dominate, learn how to use AI for preparation while strengthening trust and accountability.

Method freshness, limits, and review cycle

The method is based on task-level exposure research and O*NET-style task decomposition, with source context reviewed in June 2026. Recheck the sources and category scores as AI tools, workplace adoption, and regulation change.

Responsible use and disclaimer

This calculator is not employment, legal, financial, or career advice. Use it as a structured planning aid for learning, workflow design, and task review.

Reference method and sources

Topic Reference point Source Date Note
Task-level exposure rubric Exposure is framed as whether an LLM or LLM-powered system could reduce task time by at least 50% while preserving quality. Eloundou, Manning, Mishkin, and Rock, GPTs are GPTs August 2023 The calculator adapts this rubric into editable task scores. It estimates exposure, not layoffs or replacement timing.
Task structure O*NET organizes job data into work activities, work context, occupation-specific tasks, skills, and other descriptors. O*NET Resource Center Content Model Accessed June 2026 For a production-grade audit, map user tasks to O*NET task statements. This v1 uses user-entered task groups.
Augmentation vs automation ILO Working Paper 96 found the likely overall effect of generative AI is to augment many occupations rather than fully automate them. International Labour Organization, Generative AI and Jobs 21 August 2023 The output is intentionally framed as task transformation pressure, not an employment forecast.
US workforce exposure context The OpenAI/UPenn paper estimated around 80% of the US workforce has at least 10% of tasks exposed, and about 19% has at least 50% exposed. Eloundou, Manning, Mishkin, and Rock, GPTs are GPTs August 2023 These study-level figures are cited for context only. The page score is computed from the task mix entered by the user.

Verify the source papers and the local task score presets before treating the score as publishable.

FAQ

Does this say I will lose my job?

No. It estimates task exposure, not layoffs, wages, hiring plans, or replacement timing. A high score means parts of the task mix can likely be accelerated or redesigned with AI.

Why does it score tasks instead of job titles?

The same title can mean different work. AI exposure depends on the task mix: writing, data, support, judgment, physical work, regulated sign-off, and relationship work score differently.

What does the 50% threshold mean?

The cited OpenAI/UPenn paper labels exposure around whether LLMs or LLM-powered tools could reduce time on a task by at least half while keeping quality roughly similar.

Can I trust the exact percentage?

Use the band and task ranking more than the decimal. The exact number depends on task wording, task weights, scores, local tools, data access, regulation, and employer adoption.

What should I learn if the score is high?

Move toward the control layer: AI workflow design, prompt review, source verification, QA, exception handling, domain judgment, customer trust, governance, and metrics.

Why do regulated or relationship tasks still have some exposure?

AI can draft notes, prepare options, summarize evidence, or suggest responses. Final accountability, trust, negotiation, licensing, and live judgment still remain human-led.

Decision path

What to do next

Automation value

Measure the payback

If a task is exposed, estimate whether automating it is actually worth the cost and review burden.

 Prompt cost

Size the token budget

For exposed writing, research, or support work, estimate prompt size, context fit, and monthly API cost.

 AI workflow path

Learn the automation stack

Move from exposure screening into practical AI automation, workflow, and self-hosting guides.