Your AI Career Risk Score: What It Means If You're 40 or Older

Oct 6, 2026 · 9 min read · Career Strategy

The CareerForge Institute Editorial Team, Published by The CareerForge Institute for Mid-Career Intelligence. Institute articles carry an institutional byline and are reviewed against cited sources including O*NET, the Bureau of Labor Statistics, Eloundou et al. (2023) and the Stanford Digital Economy Lab.

If you've spent any time looking up "will AI take my job" in the past year or two, you've seen some version of a risk score. A number. A percentage. A color-coded gauge that lands somewhere between alarming and reassuring depending on which site you used.

Most of those scores are based on your job title. And for someone who's been in the workforce for 20 years, a job title tells a much smaller fraction of the story than it does for someone five years in.

Here's how a risk score is built, what it does and doesn't tell you, and why the score is only the starting point, not the answer, for professionals 40 and older.

How a job-title risk score works (and where it breaks down)

The foundational research behind AI exposure scores is Carl Benedikt Frey and Michael Osborne's 2017 paper on the susceptibility of jobs to computerisation, still one of the most-cited works in this space. Frey and Osborne estimated automation probabilities for 702 occupations, and those estimates shaped many of the risk calculators you'll find online.

The problem is that those estimates are for the occupation: the average task mix across everyone doing that job. A financial analyst two years in and a 55-year-old senior analyst who has spent 15 years building client relationships and producing judgment-based work the organization relies on get the same score if the tool uses occupation-level data.

That's not useful to either of them. It's especially not useful to the second person.

What a task-level view measures

A better-constructed risk score goes one level deeper. Instead of looking at exposure at the occupation level, it looks at tasks.

The approach: start with the specific tasks that make up the role, using the Department of Labor's O*NET database as the reference vocabulary. Classify each task as automatable (AI can handle it acceptably now or within the planning horizon), augmentable (AI changes how you do it but you're still essential), or durable (hard to automate given the nature of the work). Weight those categories by how much of a typical workweek they account for, drawing on the exposure framework Eloundou and colleagues set out in their 2023 "GPTs are GPTs" paper. Then add context from the BLS Occupational Outlook Handbook projections and the World Economic Forum's Future of Jobs Report 2025.

The career risk score here is built on that task-level logic. To be clear about what it is: the risk scores and 3-to-6-year trajectories are editorial estimates informed by that research, not values calculated directly from the Eloundou or O*NET data. Only the wage and job-growth figures come straight from BLS. The methodology page sets out exactly which parts are sourced and which are estimated.

What the score doesn't capture, and what matters more at 40+

This is where many risk calculators stop, leaving people without the right question to ask next.

The score measures exposure. It doesn't measure legibility.

These are different things, and for experienced professionals, legibility is often the more pressing problem.

Exposure is about task composition: how much of your work AI can plausibly do. Legibility is about whether your professional record communicates your value in the language current hiring systems read. You can have low exposure and still have a serious legibility problem if you haven't had to present yourself to the job market in a decade.

The Stanford Digital Economy Lab's August 2026 revision of its "Canaries in the coal mine?" study, by Brynjolfsson, Chandar and Chen, found that the measurable employment effect of AI so far is concentrated among workers aged 22 to 25 in AI-exposed occupations, whose employment has fallen roughly 19 percent relative to comparable groups. It found no comparable decline for more experienced workers, and the effect works through reduced hiring rather than layoffs.

That points to a different risk for experienced professionals. Not elevated displacement, but difficulty getting through screening when they do move: not because their skills are weak, but because their records don't describe what they contain in the vocabulary current systems expect. That is a hypothesis the RECAST framework states and is designed to test, not yet a measured finding.

RECAST calls this Legibility Debt. It accrues during stable employment: every year you spend in the same organization getting excellent at your work, your record can fall a little further behind the way that work is now described externally. It comes due at re-entry, and long single-employer tenure is where we expect it to be largest.

How to read your score

The free Career Risk Score uses three color bands: under 40 percent, 40 percent to below 65 percent, and 65 percent or higher. These are planning categories, not measured probabilities of losing your job. The tool also provides a short risk label and an explanation for your individual result.

Under 40 percent (0–39%, green band). Lower estimated disruption risk. The question to ask is whether the exposed portion, even if small, is where your career growth sits. If the automatable tasks are the ones you were promoted for, the durable ones may not carry the same trajectory.

40 percent to below 65 percent (40–64%, amber band). A signal to examine exposure in parts of your role. The question is transferability: do your durable tasks translate into a clear value argument in adjacent roles or industries, or are they specific to your current employer?

65 percent or higher (65–100%, red band). A case for starting recalibration now rather than waiting. A score of exactly 65 belongs in this band, not the middle one. Transition timelines vary widely by field and market. Starting while employed can give you more financial room and time to prepare your record for prospective employers.

In all three cases, the next step is the same: identify which of your tasks are exposed and which aren't, and express the durable ones in language that gets through current screening, automated and human, without tripping signals you didn't mean to send.

Get your free AI career risk score → Upload your resume through the score page to see your score and its explanation, then use the bands above to plan your next step.

What this platform is built to do

Most career tools are built for people entering the market or making early pivots. Their vocabulary, resume frameworks and advice are calibrated for shorter careers and more recent credentials.

CareerForge AI is built for professionals with 15 to 30 years of valuable expertise that isn't being read correctly by the systems it has to get through. The scoring, the resume analysis and the RECAST framework are all designed around that problem.

The score is where to start. It's not the answer. It's the map.

Get your free AI career risk score →, then read the methodology and the RECAST framework.