The Assumption That's Going to Cost You
A radiologist in Boston lost a major hospital contract in February 2026. Not to a cheaper radiologist. To a model that reads chest scans faster, with fewer false negatives, and at a fraction of the cost. He had 22 years of experience. He had a fellowship. He had a mortgage.
He also had a 7/10 AI displacement score. And he ignored it.
Most people approaching AI career risk are asking the wrong question. They ask: "Will AI take my job?" The real question is: "Which tasks inside my job are already gone, and what does that leave me with?"
The difference between those two questions is the difference between panic and strategy.
The scale of exposure
42% of US jobs score 7 or higher on AI displacement. That's 59.9 million jobs and $3.7 trillion in wages sitting in the high-exposure zone.
And yet. Only 3% of jobs score 9-10, meaning near-full automation. The bulk of that 42% sits at 7-8. Restructured, not eliminated. The work changes. The volume of work you're paid for shrinks. That's the actual threat, and it's subtler than people expect.
The Paradox That Breaks the Simple Story
Software developers score 8 to 9 on AI displacement. Nearly every coding task, code review, debugging workflow, and documentation process has an AI tool eating into it. By the simple narrative, software devs should be fleeing the field.
Job growth outlook: +25%.
High score. Booming demand. The paradox is real, and it's the most important thing to understand before you try to future-proof your career against AI.
Now look at medical transcriptionists. Score of 10. Job outlook of -8%. That's the danger zone. Not just high AI exposure, but a role whose remaining value disappears as the exposure climbs. No complementary skills. No expanded scope. Just substitution.
A high displacement score is not a death sentence. A high displacement score with no adjacent human value is.
Same healthcare sector. Radiologists at 7. Surgeons at 3. Nurses at 2. Same industry, three completely different futures. The title means little. The task composition means everything.
What the Data Actually Says About Risk
Andrej Karpathy published a 342-occupation analysis on March 15, 2026. The finding that got buried in the discourse: exposure score alone predicts almost nothing about job loss. What predicts job loss is exposure score combined with growth trajectory.
Here's the working framework from that analysis, applied to the data we've scored across 500+ occupations:
-
Score 9-10: disruption now. The role is being automated in real time. If you're here, you need to move, not adapt.
-
Score 7-8: restructuring in 2-3 years. Your job won't disappear. Your current version of it will. This is where most professionals are sitting right now.
-
Score 5-6: five or more years of runway. You have time. Use it to learn the tools, not avoid them.
-
Score 0-2: structural protection. Nurses at 2. Electricians at 1. Physical therapists at 3. Trades and hands-on care roles are the most durable positions in the economy right now.
The global average across all 500+ occupations is 5.3. Most people are not in crisis. Most people are in the window where preparation still works.
The counterintuitive education effect
Bachelor's degree holders average 6.7 on AI exposure. No degree holders average 4.1. Your education may be increasing your risk, not reducing it.
The Second-Order Problem Nobody's Talking About
You think your title protects you. Your title doesn't exist in isolation.
VP of Sales scores 6. Comfortable. Five-plus years of runway. But every SDR underneath that VP scores 8. When the pipeline generation layer gets automated, the VP loses the leverage that justified the role. The title survives. The organizational logic behind it erodes.
This is the second-order problem. Your own score is step one. The score of the people whose work feeds yours, and whose work you depend on to demonstrate value, is step two. Most career survival frameworks skip step two entirely.
When the work below you gets automated, the case for you gets harder to make, even if your own tasks are untouched.
81% of physicians now use AI daily. Up from 38% in 2023. That's not adoption creep. That's a structural shift in what "doing medicine" means. The physicians who thrive are the ones who redesigned their workflow around the tool, not the ones who added it on top of their existing process.
Three Actions That Actually Change Your Trajectory
Knowing your score is not enough. Here's where to put the energy, in order of leverage.
Score your tasks, not your title. Pull up your job description. List every recurring task. Assess which ones involve pattern matching, summarization, data extraction, or routine communication. Those are the tasks already in AI's lane. What remains is your actual moat.
Acquire the 56% premium. AI skills command a 56% salary premium right now. Not AI literacy. Operational proficiency with AI tools specific to your domain. A marketer who can run automated research pipelines is not in the same market as one who can't. The gap compounds every six months.
Watch the trades signal. 42% of Gen Z is now pursuing trade careers. Plumbers and HVAC technicians score 0 to 2. That's not coincidence. When the sharpest generation in recent memory pivots toward physical skills en masse, the data is already priced in. Physical dexterity, spatial judgment, and on-site adaptability are genuinely hard to automate. That's not nostalgia. That's a thesis.
The premium is real and it's growing
AI skills command a 56% salary premium today. That gap will widen before it narrows. Being early is the only way to capture the full spread.
Where do you actually stand?
500+ occupations scored 0-10. See your real exposure in 60 seconds.
The Bottom Line
The radiologist in Boston didn't lose his contract because AI got smarter. He lost it because he assumed his credentials were a moat when they were really just a head start.
The survival playbook has twelve specific moves. This article covers the first three. The full analysis, per-occupation, with detailed task breakdowns and adjacent skill maps, is what the scored report delivers. Most people who look at their score are surprised. Not always in the direction they expected.
Check where you sit before someone else's model tells you.
-
Score your tasks, not your title. The title is the container. The tasks are what AI actually targets.
-
Map your second-order exposure. The people whose work feeds yours matter as much as your own score.
-
Acquire AI skills in your domain now, not later. The 56% premium compresses as the skills commoditize.
The market doesn't care about your intent to adapt. It only registers the adaptation itself.
Find out where you stand
500+ occupations scored 0-10 on AI displacement risk. Free.