Metro train driver — tasks, one by one
The unit of analysis is the task, not the job title. Each one below carries its direction, whether the judgement rests on evidence or on platform inference, the reasoning, and what it does not establish.
Every task on this page#
Driving between stations
Automating✓ Evidence-backedSetting the train in motion, holding the speed the signalling allows, and stopping it at the right point on the platform, several hundred times a shift.
This is the oldest completed automation of a whole occupation task anywhere on this site, and it was not done with intelligence. A metro train travels a known path with no oncoming traffic, no pedestrians and no decisions about route, so the control problem is closed: the industry standard table assigns setting the train in motion and stopping the train to the machine from grade two upward, and grade two lines have been ordinary for decades. What moved the frontier recently was not better perception but cheaper retrofitting — Japan now runs commercial service where the speed between stations is entirely automatic on a line that has level crossings and no platform doors, which used to be the boundary case that required a person.
What moved is the control, not the cab. On the grade most of the world's metro kilometres actually run at, the machine drives and a person still sits in front — so the hours did not go anywhere, the content of them changed. It also says nothing about how fast this spreads, because what gates it is civil engineering: sealed track and platform doors, paid for over a decade, not software bought in a quarter.
Closing the doors
Automating≈ Platform inferenceDeciding the platform is clear and the train may leave — the last thing a driver does before the train moves, and the thing that decides whether the timetable holds.
Read the industry's own grade table as a list of five tasks and this is the fourth to move: automatic from grade four, and the driver's job at grade two and three. It is the interesting one because the difficulty was never the mechanism. A door closes itself easily; what is hard is being sure nothing is in it, and the answer was not a better sensor but a wall — platform screen doors turn an open judgement into a closed one, which is why an operator will say it cannot run unattended until every station has them.
A wall is not a machine learning to look, so nothing here transfers to any other occupation. It also does not establish that the person watching the platform disappeared: on lines with screen doors the watching moved to cameras in a control room, and the control room is staffed.
When the train stops in a tunnel
Still human-led≈ Platform inferenceA fire alarm, a person on the track, a failure between stations: deciding what to tell four hundred people, whether to move, and when to open a door onto a live railway.
The grade table puts operation in the event of disruption in the machine's column only at the top grade, and even there the disruption handled is the technical kind. Evacuation is assigned to the person on board at every grade below it, and the reason is not sentiment: it is a physical task performed once, under smoke or panic, where the cost of being wrong is not recoverable. Japan's transport ministry, listing what it must reassign before a train can run with an unlicensed attendant, puts evacuation guidance and abnormal-condition operation on that list rather than crossing them off.
Being irreplaceable in an emergency is not the same as being employed for one. These events are rare by design, and an operator can keep the capability without keeping the roster — by putting fewer people on more trains, or by keeping licensed drivers on call rather than in cabs. The protection is real and it is thin.
Being a person who is there
Still human-led✓ Evidence-backedAnswering the help point, walking the platform, dealing with the passenger who is drunk, lost, unwell or frightened — the part of the job that has no timetable.
The standard's own definition of the top grade is operation without staff on board of the trains, and the words on board are doing a great deal of work. Operators converting to it keep saying, in their own board papers, that they expect to need staff presence within the system beyond that point; and one of them, having negotiated the change with its union, recorded that station staff numbers would go up rather than down. The help points being installed on the trains are a route to a person, not a replacement for one.
Staff presence in the system is not the same job as driving a train, it usually pays differently, and it is rostered differently. Counting it as the driver surviving would be the mistake this page exists to avoid — what survived is a headcount in a company, not a seat in a cab, and the person filling it may be someone else.
Running the line from a screen
New task≈ Platform inferenceWatching every train on the line at once from the control centre, spotting the one that is behaving oddly, and deciding what to regulate before passengers notice.
This is where the work went. An operator rebuilding its system around unattended operation describes the new control centre as the hub of the new operating model and has training programmes running for its roles; a national workforce study of the same transition says the remaining focus shifts towards monitoring technology solutions and interpreting data to manage exceptions through dashboards, and names use of data and interpretation of outputs from automated systems as the emerging function. The work is real and it is a smaller number of people in one room, which is exactly what makes it worth naming rather than celebrating.
New work appearing is not new headcount, and here the arithmetic is unusually visible: one desk watches a line that used to need a driver on every train. It is also not the same career. The route from a cab to a control desk exists but it is not automatic, and the same workforce study that named this function also noted that the people in the cabs would themselves be measured by sensors on the trains.
Taking the train back by hand
New task✓ Evidence-backedDriving manually when the automation will not run: a degraded signalling mode, a section under works, moving a failed train out of the way.
The licence does not disappear, it changes shape. Japan's white paper describes an operator expanding automatic running only after first securing drivers able to drive manually in abnormal conditions, and replacing drivers with a new grade of attendant trained in about two months against the roughly nine months it takes to make a driver. That is the clearest thing anyone has published about what this automation actually does to a workforce: it does not remove the person from the front of the train, it lowers the qualification required to sit there and keeps a smaller number of licensed drivers in reserve behind them.
A reserve is not a roster. Keeping the capability requires far fewer licensed people than running the service used to, and it requires them to stay current on a skill they now rarely use, which is a training cost somebody has to carry. This also describes one country's chosen path; nothing here says other operators will take it rather than going straight to unattended operation.