■ CRITICAL RISK ■ Transportation & Logistics
Slowly but structurally, yes — not because a robot takes over the passenger seat, but because autonomous vehicles shrink the pool of people who need to learn driving at all. The job's fate is tied to the steering wheel's.
“Self-driving cars don't need lessons. They were born ready.”
Our AI replacement risk score — how we score jobs
A driving instructor's day is a loop of nervous teenagers, dual-control brake taps, and the patient repetition of mirror-signal-maneuver. The real skill is not knowing how to drive — it is reading a student's panic before it becomes a curb strike, calibrating instruction to each learner, and staying calm while someone stalls at a busy intersection for the fourth time. It is genuinely human work. The problem is that it services a skill with an expiration question mark over it.
Near term, technology is nibbling rather than biting. Driving simulators handle early-stage practice cheaply and safely; some licensing systems already use them. Apps coach learners on theory tests, and telematics can score a practice drive — smoothness, following distance, harsh braking — more objectively than a human with a clipboard. Modern cars themselves teach less: automatic transmissions, lane keeping, and parking assist mean instructors now cover fewer hard skills than a generation ago. None of this eliminates the instructor; it thins the hours each student needs.
The structural threat is demand. Robotaxis operating commercially in a growing list of cities make car ownership — and therefore licensure — optional for more young urbanites, and licensing rates among young people have been drifting down for years in several wealthy countries. If autonomy keeps scaling, driving becomes a skill like horseback riding: taught, but to fewer people, often for enthusiasm rather than necessity. Our 82 is a bet on that demand curve, not on a robot instructor. The instructors who remain will serve rural areas, commercial licenses, and anxious adults — a real market, but a shrinking one.
Automatability: our editorial assessment of current and near-term AI capability
The squeeze is already visible: simulators and driver-assist features reduce lesson hours, and youth licensing rates keep sliding in cities where alternatives exist. Through this decade, expect demand to erode market by market as robotaxi coverage expands, hitting urban instructors first. Rural instruction, commercial licensing, and remedial courses hold out longest — but a profession tied to universal driving faces a decade of steady contraction.
Most people, yes — autonomy will not be universal by then, and rural and suburban life still requires a license. But fewer people, taking fewer hours each: simulators, driver-assist features, and city dwellers skipping licensure all trim demand. The profession will exist in 2036; it will just be smaller and more specialized than today.
As a decades-long career plan, it is risky — you would be entering a market with structurally declining demand. As a five-to-ten-year income source, it can still work, especially outside major cities. If you do enter, build toward commercial licensing or fleet training rather than betting everything on teenage learners.
For early stages, largely yes: controls familiarization, hazard perception, and rehearsing dangerous scenarios are arguably better in a simulator, where a mistake costs nothing. What simulators cannot replicate is real traffic pressure and the confidence transfer of a calm human beside a terrified learner. Expect blended instruction — fewer road hours, not zero.
Specialize where automation is weakest: commercial vehicle licensing, motorcycle instruction, remedial and anxiety-focused coaching, and corporate fleet safety training. Adopt simulator and telematics tools rather than resisting them. And watch your local market — instructors in robotaxi cities will feel demand erosion years before rural colleagues do.