Will AI Replace Your Job? A Better Way to Assess Your Actual Exposure
You have read the lists. Which jobs AI will take first, which professions rank as safe, which occupations show up in red on somebody's risk chart. None of it tells you much about your own role, because the list was built from a job title, and a job title is not what determines whether your work is exposed.
Two people can share the same title, sit in the same department, and even produce the same kind of document each week, and still occupy entirely different positions when it comes to actual risk. One of them does something that would be genuinely difficult to reproduce without them. The other does something that could, in time, be reproduced fairly easily. The title gives you no way to tell which is which. Only the work does.
So the more useful question is not whether your occupation makes anyone's endangered list. It is whether what you specifically do, week to week, is structurally replaceable or structurally expandable. Those are different things, and the difference is observable if you know where to look.
Two kinds of exposure, not one scale of risk
A replaceable role is one where most of what you produce follows a documented procedure. The skill involved is mostly execution speed, and the experience behind it is mostly familiarity with the process rather than judgement about it. When work like this gets evaluated, whether by a manager, a restructuring committee or simply the passage of time, the test applied is blunt: can something like this be produced for less, somewhere else? If the answer is yes, the role contracts. Tasks migrate. Scope narrows. Authority stays where it was, or shrinks further.
An expandable role behaves differently. It is one where removing the routine parts of the job actually creates room for more consequential contribution, because the person already sits somewhere judgement matters. The test here is not about cost. It is about what becomes possible if this person's scope grows. Automation strips out the procedural layer and what is left underneath gets more room to operate, not less.
This is not a matter of seniority. A senior analyst who spends most of their time producing standardised reports can occupy a replaceable position despite the title. A mid-level coordinator who reliably spots a dependency conflict before it cascades into three delayed projects can occupy an expandable one. Where you sit in the decision chain matters more than where you sit on the org chart.
Where the value in your role actually sits
Most job descriptions blur this distinction, because they list responsibilities rather than consequence. Vendor management, process documentation, team coordination: all real, all necessary, and all silent on the question of what would actually break if you stopped doing them.
It helps to separate what you produce into two rough categories. Some of your output is transactional. It keeps things running, but its absence would barely register beyond a short delay: a checklist completed, data logged, information forwarded to the right people. Other output is consequential. It changes what a stakeholder decides, how a risk gets handled, or which direction a project takes. The same person often produces both kinds in the same week, and the proportion between them, not the job title, is the real indicator of exposure.
There is a fairly direct way to test which category a given piece of your work falls into. Imagine you were absent for a month. If the work continued at roughly the same pace and quality with someone temporarily covering, that part of your role is replaceable. If decisions stalled, if a particular kind of risk went unnoticed for longer than usual, if people started escalating things because the person who normally caught them wasn't there, that part is expandable. It is a genuinely different category of contribution, and it usually only becomes visible in its absence.
A second test looks at how people respond to what you hand them. If they receive your output, note it, and move on without changing anything about what they do next, you are delivering information. If they adjust a decision, reprioritise something, or escalate differently because of what you gave them, you are delivering leverage. Information can usually be produced elsewhere. Leverage is harder to relocate, because it depends on judgement that has already proven itself against real outcomes.
What to actually check before you draw a conclusion
Run this against your own work honestly rather than optimistically. Ask how much of what you produce follows a repeatable procedure that someone could learn from documentation alone, versus how much requires you to weigh trade-offs that aren't written down anywhere. Ask whether your involvement in decisions has been growing, staying flat, or quietly narrowing over the past year, not just this month. Ask what specifically would degrade if you disappeared for four weeks: nothing much, a delay, or a decision made with worse information. And ask whether the requests reaching you have been shifting from "here's a problem, what should we do" towards "here are the instructions, please execute."
None of these questions gives you a verdict on their own. A single quiet month proves very little. A single meeting where nobody asked your opinion could mean anything. What matters is the pattern across several of these signals over time, not any one data point in isolation.
It is also worth resisting the urge to sort your entire role into one bucket or the other. Almost nobody's job is purely replaceable or purely expandable. Most roles are a mixture, and the honest exercise is figuring out roughly where the balance currently sits and which direction it has been moving, rather than searching for a single verdict that settles the matter for good.
None of this produces reassurance, and it isn't supposed to. What it produces is something more useful than reassurance: a reasonably accurate picture of where your contribution currently stands, which is the only sensible starting point for deciding what to redesign, protect, or build further, before the decision gets made for you.
David Taylor
Interested in going further?
If this diagnostic raised more questions than it answered about your own role, The AI-Ready Career sets out how to systematically redesign your contribution once you know where the exposure actually sits.
Ad · Amazon affiliate link.
