What 23,000+ working people actually report
Is my job safe from AI?
By Jon Miksis, founder of Make the Leap · Data updated October 2026
Every week another headline announces the end of some profession, and every week actual working people have to calibrate: is my job on the list, and what is the smart move if it is? We have an unusual window into that question - our assessment asks working adults directly how AI is landing on their work, and enough of them have answered to see the pattern.
The honest headline from our data: about one in ten working people say AI makes them a little nervous - but the average hides the spread. Retail, administrative, and healthcare workers report two to four times the concern of executives. And the people who feel nervous are not naive; the likeliest explanation in our read is that they are the ones seeing automation reach visible parts of their own week first.
The contrast with the headline cycle is stark enough to state plainly: in our dataset of 23,183 people actively exploring a career change, just 10% said AI makes them a little nervous - while burnout, named by 49%, is the pain that actually pushes people to move. That is a self-selected career-changer sample, not a representative workforce poll. But among the people worried enough about their working lives to act, AI is rarely the reason - exhaustion is.
Is my job safe from AI? A 12-question check
Occupation-level risk calculators average over everyone who shares your job title. These twelve yes/no questions ask about your week instead - the four properties that make work hard to automate (judgment under ambiguity, trust-based relationships, hands-on skilled work, accountability for outcomes) plus the task-versus-judgment hour split. Your answers stay in your browser, and the result is a count of the protective signals you reported - not a prediction about your job.
0 of 12 answered
1. Judgment under ambiguity
In a normal week, do you make calls where the right answer is not written down anywhere - where you are weighing incomplete information?
2. Judgment under ambiguity
When something goes wrong, are you the person who decides what happens next, rather than the person who reports it upward?
3. Judgment under ambiguity
Does your work regularly need context that lives in your head or in the room, rather than in a system that could be read?
4. Trust-based relationships
Do people ask for you by name, rather than for whatever your role happens to produce?
5. Trust-based relationships
Does a meaningful part of your work depend on a relationship built over months or years?
6. Trust-based relationships
Is part of what you are paid for the fact that a person - not an answer - is present for it?
7. Hands-on skilled work
Does your work require you to be physically present somewhere that is meaningfully different each time?
8. Hands-on skilled work
Do you assess, diagnose, or fix things whose real condition can only be judged in person?
9. Accountability for outcomes
Is your name on the result - legally, financially, professionally, or reputationally?
10. Accountability for outcomes
If nobody held your specific accountability, would that be a real problem for your employer or clients?
11. The hour audit
If you sorted last week's hours, would more than half be deciding, persuading, or owning outcomes - rather than processing, drafting, scheduling, or routine analysis?
12. The hour audit
If the routine parts of your job were automated tomorrow, would there still be a whole job left?
What our data shows about people nervous about AI
The most common pains this group names: burnout (51%), feeling underpaid (46%), a lack of meaning (35%), a bad culture (30%), not using their strengths (30%).
AI worry, by profession
Share of each profession's assessment takers who say AI makes them a little nervous:
Where the data points next
Where the assessments of people nervous about AI point: notice how many themes involve owning outcomes rather than performing tasks.
The pattern across our data: task-shaped work (scheduling, drafting, processing) is automating; judgment-shaped work (deciding, persuading, owning outcomes) is not. The people who feel nervous about AI and act on it move up the judgment stack or into hands-on work - the lanes AI has reached slowest so far.
The selection-side companion to this page - which careers sit on the four moats AI has reached slowest, and how to move toward them without starting over - is our AI-proof careers guide. If you want those careers ordered - by how many of the four moats each one stacks, and what would move each down the list - jobs AI can't replace, ranked is the list-shaped third page of the same hub.
What people nervous about AI in our data earn today
Current household income bands this group reports - useful for calibrating what a transition has to protect:
What independent research says about AI job exposure
Our data measures how workers feel about AI - how many say it makes them nervous. That is a different thing from predicting which jobs will disappear - so it is worth knowing that independent task-level research points the same direction.
The International Labour Organization's global index estimates that roughly one in four workers are in occupations with some generative-AI exposure, but concludes that transformation is far more likely than wholesale replacement, because most occupations still contain tasks that require human input - and its 2026 research brief explicitly warns against reading exposure measures as job-loss forecasts. The OECD's AI Exposure Measure lands on the same axis our self-audit uses: current AI capability sits closest to routine information processing and codifiable tasks, and furthest from work built on contextual judgment, interpersonal understanding, complex decisions, and responsibility.
That is why the audit below starts with your actual tasks, not your job title.
- ILO - Generative AI and Jobs: A Refined Global Index of Occupational Exposure (2025)
- OECD - The OECD AI Exposure Measure
How at risk is my job from AI? A five-minute self-audit
Occupation-level risk calculators average over everyone who shares your title. Your actual exposure lives in how your specific week splits between task-shaped and judgment-shaped hours. Score yourself honestly:
- 1
Count your task-shaped hours - processing, drafting, scheduling, formatting, moving information between systems. These automate first. If more than half your week lives here, that is real exposure.
- 2
Count your judgment-shaped hours - deciding under ambiguity, persuading, owning a budget, a number, or an outcome. These are the hours AI extends rather than replaces.
- 3
Check who owns the relationships. If clients, patients, or stakeholders would follow you out the door, you hold the layer that automates last.
- 4
Check whether your output gets signed off or simply trusted. Work someone must verify automates slower than work nobody double-checks.
- 5
Watch where your own tools improved this year. The parts of your job AI already drafts for you are the parts an employer eventually stops paying full price for.
If the audit comes back mostly task-shaped with replaceable relationships - the honest version of 'is my job screwed' - the answer is usually not fleeing to a random 'safe' job. It is repositioning toward the judgment layer of work you already know. Our free assessment reads your actual skills and experience and maps which of your strengths carry that move.
Start this week
Three income-safe first moves. None require quitting, announcing, or being ready:
- 1
Audit one normal week: mark each recurring hour task-shaped (automatable) or judgment-shaped (defensible). The ratio is your actual exposure, headline-free.
- 2
Automate one of your own task-shaped hours this month with an AI tool. The person operating and improving the automation holds a stronger position than the person whose role is defined by the task it replaces. Being the operator is not permanent immunity - but it is the right side of the divide to stand on.
- 3
Pick one judgment asset to grow this quarter: own a budget, a client relationship, or a decision. Accountability is the moat.
Honest answers
Which jobs does our data show are most nervous about AI?
Retail and hospitality workers report the highest concern in our assessment data - about one in five - followed by healthcare workers, nurses, and administrative workers; executives report the lowest. The full by-profession table is on this page and in our research hub, computed from each profession's own respondents.
Should I change careers because of AI?
Change because the work is wrong for you; reposition because of AI. For most people the right move is not fleeing the field but shifting weight inside it - from the automatable task layer to the judgment layer. If the field itself is task-shaped, that is when a real change is rational.
What skills are actually AI-resistant?
The recurring winners across our data's path themes: judgment under ambiguity, trust-based relationships, hands-on skilled work, and accountability for outcomes. Note that 'creative' is not automatically safe - creative production automates faster than creative judgment.
What's the fastest way to lower my AI exposure without changing careers?
Audit your own week: which hours are task-shaped (automatable) and which are judgment-shaped (defensible)? Grow the second deliberately - own budgets, decisions, client relationships. Our assessment maps your strengths against exactly that split.
How do I know if AI will replace my job?
Ignore the headlines and audit your own week: hours that are task-shaped - processing, drafting, scheduling - are automatable, while hours that are judgment-shaped - deciding, persuading, owning outcomes - are defensible. Your personal ratio tells you more than any occupation-level prediction, because jobs rarely disappear whole. Tasks do.
What should I do if AI takes my job?
Stabilize first: runway, benefits, filings. Then move up or sideways - up the judgment stack in your own field, or sideways into hands-on and relationship-heavy work that AI has reached slowest. Your experience does not evaporate with the role; the tasks did. The assessment maps which of your strengths are the durable kind.
Which jobs are safest from AI?
In our worry data, executives report the lowest concern; retail, administrative, and healthcare workers the highest. But 'safe job' is the wrong unit - tasks automate, not whole jobs. The durable pattern across our data's path themes: judgment under ambiguity, owned outcomes, hands-on skill, and trust-based relationships. Our AI-proof careers guide maps roles built on exactly those four.
Will learning AI skills protect my job?
Partially. Being the person who operates the tools beats being the person the tools replace, and in our data the professions with the highest AI worry are exactly where that move is most available. But tool skills commoditize fast - pair them with judgment, domain depth, and owned outcomes for real protection.
See what the data says about you.
The free assessment reads your actual answers - your strengths, constraints, and the pattern keeping you stuck - and gives you personalized paths in about 10 minutes.
Take the free Career LeapKeep reading
More situations like yours
Numbers on this page were computed on 2026-10-07 from 28,376 completed Make the Leap career assessments. Cohort: assessment takers who said AI makes them a little nervous. The by-profession table uses each profession's own respondents. Percentages use the respondents who answered each question; path themes are counted once per person from their personalized assessment paths. Full dataset and methodology: our research hub.