The Bottom Rung Is Disappearing. The Fire Service Is Standing On It.
Robert Grand · Battalion Chief who still runs calls
Entry-level hiring in white-collar America is collapsing. Postings for entry-level jobs are down roughly 35 percent since January 2023, and in tech it’s closer to a cliff, with one market report putting entry-level hiring down 73 percent year over year (The Crisis of Entry-Level Labor in the Age of AI (https://www.rezi.ai/posts/entry-level-jobs-and-ai-2026-report)). One in five companies has frozen entry-level hiring specifically because of AI, and nearly half expect to eliminate entry-level roles entirely by 2027. The business press is covering it as a jobs story: a brutal market for new graduates, colleges scrambling to fill an experience gap they didn’t create (AI is wiping out entry-level jobs (https://fortune.com/2026/05/15/ai-entry-level-jobs-higher-education-experience-gap/)).
Read the same numbers from a fire station and it’s not a jobs story at all.
They just automated away the thing we call probation!
What the Hiring Numbers Are Actually Counting
Here’s what those industries are discovering the hard way. The junior work AI now does was never just cheap labor. It was the training ground. The first-year analyst grinding through spreadsheets, the junior developer fixing small bugs, the new associate summarizing case law: none of that work was glamorous, and all of it was load-bearing. Somewhere in those thousands of unremarkable tasks, judgment formed. Eliminate the routine work and you eliminate the reps. Ten years later you have nobody seasoned enough to supervise the AI that replaced them.
The engineering world is already saying this out loud. More than half of engineering leaders plan to hire fewer junior developers, and the same leaders are now asking who trains the next generation when the bottom rung is gone (Junior Developers in the Age of AI (https://www.softwareseni.com/junior-developers-in-the-age-of-ai-who-trains-the-next-generation-of-engineers/)).
And there’s a harder number underneath the hiring data. Research from Anthropic found a 17-point comprehension gap when juniors learn with AI assistance: about 50 percent understanding of the work, versus 67 percent without the assist. The tool hands you the answer without the struggle, and the struggle was where the learning lived. That’s not an argument against the tools. It’s a fact about how expertise gets built, and it doesn’t care how good the tools get.
Judgment Is Built on Four Hundred Boring Calls
The fire service runs the apprenticeship model these industries are now scrambling to rebuild. We never stopped. The rookie rides backwards, runs the routine calls, pulls hose, does the inspections nobody wants, taps out narratives in the parking lot, and somewhere in those thousands of reps becomes the engineer, then the officer, then the one everybody looks at when the call goes sideways.
Nobody develops clinical judgment on the cardiac arrest. They develop it on the four hundred routine transports before it. The lift assists, the abdominal pains, the little old lady who ‘just doesn’t feel right’ and turns out to be the sepsis save because a second-year medic had seen enough almost-normal to recognize not-normal. The slow calls are where the learning happens. Every veteran in the service knows this, even if nobody wrote it down, because nobody had to.
For decades that model looked old-fashioned next to the credential-driven white-collar career path. Four years of college and a desk versus riding backwards and washing the apparatus. Now the desk jobs have automated away their own bottom rung and discovered the rung was holding up the whole ladder. For once, Fire & EMS is ahead of the curve by standing still.
That’s the good news. Here’s the uncomfortable part.
No Single Tool Will Look Like the One That Broke It
AI can’t take the seat on the engine. It can’t pull a line, can’t package a patient, can’t make entry. The physical reps are safe. But the cognitive reps are not, and the cognitive reps are half the apprenticeship.
AI narrative writing takes the charting reps. An AI triage assist takes the differential-building reps. AI-generated preplans take the get-in-the-building-and-look reps. Size-up support takes the read-the-smoke reps. Each one of those tools is genuinely useful, and each one absorbs a piece of the routine cognitive work that used to build judgment in new firefighters and medics. Every adoption decision will look reasonable on its own. None of them will look like the one that broke the pipeline.
Each tool feels like efficiency. Together they’re a withdrawal from an account nobody is reconciling.
I’ll be honest about my own side of this. I use AI every day and write about it multiple times a month, and the tools have made me better, not worse. I write more now because the AI sparks ideas and helps me connect dots I would have walked past on my own. But I can feel the trade underneath it: when a model drafts something I used to build from scratch, the output is faster and my own recall is thinner. I’m 24 years in; my judgment account is full. The probationary firefighter who starts next month doesn’t have that cushion, and the tools will be in front of them from day one.
Flight Programs Already Solved This. They Call It Stick Time.
Aviation has lived with this exact problem for forty years. Autopilot can fly the plane better than the pilot for most of the flight, and flight programs still force pilots to hand-fly approaches, because the day the automation drops out is not the day to discover your hands forgot. Stick time is protected as a matter of doctrine, not nostalgia.
That’s the template for Fire & EMS training divisions. Reps become a protected resource. Somebody decides, deliberately and in writing, which tasks a probationary member still does manually even when the tool does it faster: which narratives get written without the assist, which preplans get walked instead of generated, which differentials get built out loud in the back of the ambulance before anyone looks at the screen. Then it goes into the taskbook and the field training program, so it survives the next budget cycle and the next staffing crunch.
Because the pressure will always run the other way. When you’re short-staffed, the argument for letting the tool do the routine work is unanswerable in the moment. Protecting developmental reps looks like inefficiency on every individual shift and only shows its value years later, which makes it exactly the kind of investment that loses budget fights. It takes a training officer, a medical director, and a fire department leader willing to defend slow on purpose.
The Recruiting Window Nobody Has Named
There’s a second opportunity sitting inside this story, and it’s a recruiting pitch. The white-collar career ladder is broken at the bottom. A 22-year-old who did everything right is staring at entry-level postings down 35 percent and a market where almost half of employers plan to eliminate those roles entirely. The trades and public safety are among the few places left in the country where that same 22-year-old can still climb: a real apprenticeship, a defined path from the back seat to the front seat to the command vehicle, and a pension at the end of it.
Fire & EMS has spent years competing for candidates against private-sector paychecks and losing. The private sector just orphaned an entire generation of entry-level talent. The departments that tell the ladder story now, in their recruiting materials and their academy pitches and their high school career fairs, get first pick of candidates the rest of the economy left standing outside.
The Ladder Is the Asset
By 2030 the industries that automated their junior work will be facing a seniority gap they can’t hire their way out of, because every competitor has the same hole. Fire & EMS can be the counter example: the profession that adopted the tools and protected the pipeline, on purpose, with the reps written into doctrine. Or it can sleepwalk into the same gap on a five-year delay and find out in 2031 that its new lieutenants never built the judgment the job assumes.
The apprenticeship was never the old-fashioned part of this profession. It was the part everyone else is about to spend a fortune trying to rebuild.
Robert Grand is a Battalion Chief at Eugene Springfield Fire with 24 years of service. He writes Frontline Intelligence, a newsletter on operational doctrine, technology, and leadership in Fire & EMS.
Subscribe
From the floor, not the vendor booth. Two times a week.