Why most of this job stays in the bay
Fixing cars is physical work with a diagnosis attached. A technician hears a noise on a road test, puts the car on a lift, feels for play in a joint, then gets a hand into a space built for parts and not for people. Software can read a fault code. It cannot reach behind an engine, break a rusted bolt loose, or tell by feel that a bushing is shot.
The second reason is variation. A ten-year-old pickup with rust, a leased sedan under warranty and a new EV with a sealed battery pack all arrive at the same shop on the same morning. Each needs a different plan, different access and a different conversation about cost. Jobs like replacing brake assemblies, chasing an intermittent electrical fault or test driving a car after the repair all sit with a person for now.
There is a third reason that gets less attention: responsibility. Someone signs off that the vehicle is safe to drive away. That accountability, and the customer conversation behind it, is part of the work. Our scoring method weighs that task mix rather than the job title.
What AI does, what it assists with, and what it leaves alone
Start with the share of task time AI can handle on its own: 0%. That slice is paperwork and lookup, not repair. Think of turning scan-tool codes into a plain-language summary, drafting the repair order and estimate, and searching service bulletins and wiring diagrams faster than flipping through a manual. These are the tasks where a text tool genuinely removes minutes from the day. The coverage score explained page shows how that share is built.
Next, the assisted share: 7%. Here AI narrows a guess but does not close it. Symptom-plus-sensor-data tools can suggest likely causes before you lift the hood, and parts lookup now suggests the right component and labor time for the model in front of you. A technician still confirms it on the vehicle, because the suggestion is only as good as the data the car gave up.
Everything else needs a person: 93%. That is the lift work, the road test, the teardown, the calibration after a windshield or sensor is touched, and the explanation to an owner deciding between a repair and a trade-in.
What has actually been tested
Not much, and that matters. The parity evidence grade for this job is D, which on our scale means no study has measured an AI system against a qualified technician on this occupation’s own tasks. So we publish no parity number here, and you should be skeptical of anyone who does.
A settled answer would need a specific test: common faults on real vehicles, an AI-guided setup against trained technicians, with fix rates, comeback rates and time per job recorded and the method published. Chat demos and single repair videos are not that. Until that exists, the honest read is task erosion in diagnosis and paperwork, not a job in retreat. The full scoring method sets out how we grade evidence from A to D.
The labor data points the same way. The Bureau of Labor Statistics counts about 704,640 automotive service technicians and mechanics in the United States, with median pay near $50,620 and projected employment growth of about 5% between 2025 and 2035 (BLS, 2025).
When the picture could change
Most likely after 2045 (8 in 10 of our scenarios). Our replacement year method explains what that window is measuring.
Two things could pull it earlier. First, dexterous robots that work reliably in a cluttered bay, not a clean factory cell, since most of this job is physical handling. Second, richer vehicle data: as more cars stream fault and wear data off-board, more of the diagnostic step can happen before the car arrives.
Two things hold it back. Access and condition: rust, crash damage, aftermarket parts and awkward packaging defeat tidy automation. And cost with liability: a robot capable of this work carries a capital bill and a safety case, while a trained technician is hired today and insured today. Shops with two lifts do not buy humanoids. Our guide to humanoid robots and physical jobs covers why that step is slow.
How to stay needed as a technician
Lean into the parts of the work that stay human. Three worth protecting: hands-on diagnosis that starts with a road test and your own senses, complex teardown and reassembly where access is the hard part, and the customer conversation where you explain what is urgent, what can wait and why.
Two skills raise your floor. High-voltage and EV service work, including battery and thermal systems, is where demand is moving. Advanced driver assistance calibration is the other: cameras and radar need a trained setup after glass, suspension or body work, and that is billable, documented and growing.
What to do: treat AI scan summaries as a first hypothesis and write down how often they match what you actually find.
If you are weighing a move, look at nearby trades first. Bus and truck mechanics and diesel engine specialists work on heavier equipment with steadier fleet contracts. Automotive body and related repairers and automotive glass installers and repairers both now carry sensor calibration work that did not exist a decade ago. You can put any two of them side by side on our compare any two jobs tool.
For the wider picture, the vehicle and mobile equipment repair family shows how neighboring roles score, and the auto repair sector page tracks the industry around the shop. If you want the broader shortlist, see the jobs that mostly need a person list.