Will AI replace civil engineers? The honest answer

The fear assumes engineering is a fixed pile of work

Take the fear in its strongest form: a machine that produces the deliverable end to end, and an engineer with nothing left to do. It deserves an honest answer, which means answering the assumption underneath it. The assumption is that there is a set amount of engineering to be done, so whatever the machine does is subtracted from the engineer's share. Economists named this mistake a century ago, the lump-of-labour fallacy, and it is the same one made about every machine since the loom. It has been wrong every time, for one reason: what people want done is not fixed, and cheaper capability expands it.

We build one of these tools, so weigh our optimism accordingly. But the argument does not rest on our word. It rests on what has happened every time a kind of work got cheaper: the output gets cheaper, more of it is wanted, and the wanting creates more work, not less. The composition shifts, sometimes hard. The total does not shrink.

The ATM was supposed to end the teller. It did the opposite.

The clearest case is the cash machine. When ATMs spread across America in the 1980s and 90s, the teller looked finished: a machine now did the one task a teller was for. Instead the number of tellers rose, from about half a million toward six hundred thousand. The machines made a branch cheaper to run, so banks opened far more branches, and each branch still needed people, now for the parts a machine could not do: sorting out the customer who was stuck, selling the mortgage, holding the relationship. The task was automated. The job grew, and moved up.

This is the pattern, not the exception. A job is a bundle of tasks; automate some, and the ones left over tend to become more valuable, not less. Civil engineering has lived a smaller version already. The drawing board did not survive CAD; the engineer did, and used the tool to take on more, and more complex, work than the board ever allowed. The tracing was automated. The engineering grew around it.

The tasks that stay become the valuable ones

Apply that to a project. When the machine produces the deliverable, the tasks that do not automate are the ones that were always worth most: deciding what the study must show, judging whether a model fits this ground, reading a site the documents do not capture, putting a name to a recommendation others will build on. They used to sit buried under production hours. Now they are the job. The engineer is not doing less engineering; they are doing the concentrated kind, and it is worth more.

The people who run large engineering organisations describe the same shift from the top.

The work was never going to get done otherwise

So the honest answer is not that your job is safe. It is that there is far more engineering worth doing than the profession can currently reach. The infrastructure backlog, the deferred maintenance, the studies never commissioned because the fee could not carry them, the alternatives dropped because evaluating them cost too much: that is the fixed-pile assumption meeting reality. Cheaper engineering does not empty a queue of engineers. It starts clearing a queue of work that was never getting done.

The labour numbers point the same way. Cedefop, the EU's vocational-training agency, projects, and a projection is all it is, that around 4.1 million people will be needed in European construction between 2022 and 2035 just to replace those leaving the occupation, mostly to retirement, while most European markets already report skills shortages. Sweden's own delivery agency has said as much about its national transport plan, which we covered in The plan is funded. The engineers aren't there. The sector cannot hire its way through the work it has already taken on.

"It is like when the personal computer came out in the mid-90s, and everyone wanted one but nobody knew how to use it," says Kenneth, an engineering director. The capability arrives before the working method. The engineers who learn to direct it pull ahead of the ones who do not: not machine against engineer, but the engineer with an AI colleague against the one without. Replacement was never the real threat. The waste is the engineering that never gets done, and that is the pile that finally starts to clear. Not fewer engineers — more of what engineering is for.