A radiologist in Boston lost a hospital contract last fall. Not to a cheaper radiologist. To software running on a server in Dublin. His read accuracy was 97%. The algorithm hit 99.1% on the same test set. Two years earlier, he'd been told his specialty was "too complex for automation."
His AI Displacement Score: 7 out of 10.
That number exists for 500+ occupations now. And most people are reading it wrong.
The question everyone asks is "will AI replace my job?" The more useful question is when. Because the timeline by score is specific. And specific means actionable.
What Most People Get Wrong About AI Job Risk
People assume the danger zone is obvious. Repetitive work. Factory floors. Data entry. They think a knowledge worker with a graduate degree is safe. The data says the opposite.
The salary paradox
Jobs paying $100K+ average a 6.7 AI exposure score. Jobs under $35K average 3.4. Higher education, higher risk. The data doesn't care about your credentials.
Bachelor's degree holders average 6.7. No degree: 4.1. This isn't a rounding error. It's a structural shift. Knowledge work, the stuff you went to school for, turns out to be exactly what language models are trained to replicate.
Meanwhile, 42% of Gen Z is pursuing trades. Plumbers score 1. HVAC technicians score 2. They figured something out that MBAs haven't yet.
The physical world is the last moat. Not expertise. Not credentials. The fact that your hands have to be in the room.
Nurses score 2. Electricians score 1. Physical therapists score 3. These jobs require bodies, judgment in unpredictable environments, and tactile feedback no model can replicate today. The global average across all 500+ occupations is 5.3. Half the working world is closer to disruption than safety.
The Timeline by Score: When Disruption Actually Hits
This is where it gets specific. The AI job displacement timeline isn't one wave. It's three, arriving at different speeds.
-
Score 9-10: Disruption is now. Only 3% of occupations land here, but they're already contracting. Medical transcriptionists: score 10, job outlook -8%. Data entry clerks, insurance underwriters, basic paralegal work. The wave didn't warn them. It just arrived.
-
Score 7-8: 2-3 years. This is where 42% of US jobs sit. 59.9 million workers. $3.7 trillion in wages. These jobs won't disappear overnight. They'll restructure. Fewer people doing more. The ones who stay will need to justify their floor space against the algorithm's cost.
-
Score 5-6: 5+ years. Meaningful exposure but time to adapt. These roles have enough complexity, judgment, or physical component that full restructuring requires more than current AI can deliver. The window is open. Not forever.
But here's where the AI automation timeline gets complicated. A high score doesn't mean contraction. It can mean transformation.
The software developer paradox
Software developers score 8-9 on AI exposure. Job outlook: +25% growth. High disruption score. Booming demand. The job is restructuring around AI, not disappearing into it.
That contrast is the real lesson of the AI automation timeline. Score measures exposure to restructuring. It doesn't determine whether that restructuring kills headcount or multiplies output. The difference is whether you're the one wielding the tool, or the one being replaced by it.
Same Industry. Completely Different Futures.
The most dangerous assumption is that an industry protects you. It doesn't. The score lives at the task level, not the sector level.
Consider healthcare. One industry. Two completely different stories.
-
Surgeons: score 3. Hands in the body. Judgment under pressure. Physical dexterity in unpredictable anatomical conditions. AI assists but doesn't replace the person holding the scalpel.
-
Radiologists: score 7. Pattern recognition on images. Structured output. Exactly what models are built for. 81% of physicians now use AI daily, up from 38% in 2023. That adoption curve is compressing the timeline fast.
The second-order effects are just as important. VP of Sales scores 6. Manageable exposure. But the SDRs under that VP? Score 8. When the AI automation timeline hits the people below you, your own role contracts whether your score says so or not. Fewer SDRs means less pipeline means fewer VPs needed to manage it.
Your score doesn't exist in isolation. It exists inside an org chart. When the people below you get automated, you feel the pressure too.
Andrej Karpathy's March 2026 analysis of 342 occupations confirmed what the score data shows: the vulnerable tier isn't entry-level work. It's mid-layer knowledge work. The coordination, synthesis, and communication roles that sit between execution and strategy. That layer is thin. And it's thinning.
What the Data Actually Tells You to Do
Knowing your score without a plan is just anxiety with a number attached. The AI job displacement timeline is useful only if it changes behavior.
The skills premium
AI skills command a 56% salary premium right now. The market is already pricing in who can use the tool versus who gets replaced by it. That gap will widen.
Know your actual score, not your assumed score. Most people underestimate their exposure by 2-3 points. The score lives at the task level, not the job title. A "manager" who spends 80% of their day on synthesis and reporting is far more exposed than their title suggests.
Map the tasks inside your role, not just the role. Which parts of your day could be done by a capable language model with the right prompt? That's your real exposure surface. The answer is usually higher than comfortable.
Move toward the physical, the relational, or the strategic. The durable jobs share one trait: they require something a model can't do alone. Physical presence. Trust built over time. Judgment with incomplete information and real stakes. The closer your work is to those elements, the longer your runway.
The 56% salary premium on AI skills isn't a bonus. It's a signal. The market is already sorting into two groups. People who wield AI as a force multiplier, and people whose roles are being multiplied away.
Which side of that line you end up on isn't determined by your industry or your degree. It's determined by whether you acted when the timeline was still yours to control.
What's your exact score?
500+ occupations scored 0-10. Free lookup. Takes 60 seconds.
The Danger Zone Is Specific
A score of 9-10 with a declining job outlook. That combination is the real signal. Not high score alone. High score plus contracting demand.
Medical transcriptionists sit at score 10 with -8% job outlook. That's not a trend. That's an exit ramp closing. The medical transcriptionist in 2019 who was told "AI can't handle medical terminology" is now the cautionary data point in this article.
Compare that to software developers. Score 8-9, job outlook +25%. Same high AI exposure. Opposite trajectory. Because software developers are using AI to expand what they can build, not watching AI displace what they were hired to do.
The difference is agency. One group is subject to the automation timeline. The other is setting it.
That's the frame that matters when you're reading your own score. Not "is this number high?" but "am I the person using AI to do more, or the person AI is being used to replace?"
The full picture, including the 12 specific actions mapped to each score tier, lives in the survival report. This article gives you the framework. The report gives you the playbook for your exact occupation.
Bottom Line
The AI job displacement timeline isn't coming. For 3% of occupations, it already arrived. For 42% of US jobs, it's 2-3 years out. The score tells you which group you're in. What you do with that information is the only part still up to you.
Find out where you stand
500+ occupations scored 0-10 on AI displacement risk. Free.