Job Resiliens

Research  ·  AI Job Risk Index  ·  2026 Edition

The AI Job Risk Index 2026

88 occupations, each researched individually — what's automated, what isn't, and why — sorted into three exposure tiers. Not a survey, not a projection: a synthesis of our own published, task-by-task research.

Published August 2026 · Job Resiliens Research · Updated as occupations are added or re-researched

How to read this Each tier reflects our own editorial research verdict for that occupation, documented on its own page with reasoning, not a numeric score from a survey. "High exposure" means the role's core tasks are substantially automatable today; it does not mean the job disappears — most roles unbundle into an automating part and a durable part. Click through to any occupation for the full task-by-task breakdown.

High exposure — core tasks are substantially automatable today

Routine, structured, high-volume tasks with a checkable right answer dominate these roles. Two of these — transcription and general translation — are furthest along, with most of the routine work already handled by AI. The rest still have a real, durable half (advisory, brand, exceptions, judgment) worth building toward.

Moderate exposure — the role unbundles, it doesn't disappear

The largest tier by far. These roles split cleanly into an automating half (execution, standard cases, production volume) and a growing half (strategy, complex judgment, client relationships) — and the split ratio varies a lot person to person within the same title.

Low exposure — structurally protected, not just "safe for now"

What holds this tier together isn't the industry, it's the structure of the work: physical presence, a license carrying personal liability, or trust built with a specific vulnerable person. None of that changes with a better model — it's a feature of how the work and its regulation are built, not a temporary gap in AI capability.

Methodology, briefly

Each occupation above was researched individually: what AI tools can already do in that role today, what remains genuinely out of reach (and the structural reason why — license, physical presence, trust, liability), and the honest near-term trajectory. Tiers are our own editorial synthesis of that research, not a numeric average of external surveys. For the full reasoning behind any single role, follow its link above — every page documents its own publish date and reasoning. For how this connects to your own personalized AI Exposure Score, see how we score AI resilience.

Cite this research

This index is free to cite and link to. If you're writing about it, the reference below is ready to use:

Suggested citation Job Resiliens. "The AI Job Risk Index 2026." Published August 2026. https://jobresiliens.com/research/ai-job-risk-index/

For a specific occupation's figures, cite that occupation's own page instead — e.g. Job Resiliens, "Will AI Replace Accountants?", jobresiliens.com/ai-risk/accountant/ — since that's where the underlying task-by-task reasoning actually lives.

Embed this chart

Want to embed the exposure-tier chart on your own site or newsletter, with attribution? Paste this snippet — it stays live and updates automatically as we add or re-research occupations:

<iframe src="https://jobresiliens.com/research/ai-job-risk-index/embed/" width="100%" height="360" style="border:1px solid #DDE3ED;border-radius:12px;" loading="lazy" title="Job Resiliens AI Job Risk Index 2026"></iframe>

Get your personalized AI exposure score

This index is about roles in general. For your specific responsibilities: a free, task-by-task AI job risk score plus an AI reskilling plan based on your resume.

Get your AI Exposure Score