Skip to content
man bot
Compare

BlogWork that is forming

Six job ads from 2040

A bit of fiction, clearly labelled. Six job adverts for roles that barely exist yet, each one grown from something already happening today, with a note on the real trend behind it.

Work that is forming 5 min read

Everything below is fiction. Nobody is hiring for these jobs today. But every one of them is grown from something real, and after each ad there’s a short note on the trend it comes from. Think of it as a postcard from a future that could plausibly turn up.

1. Robot fleet supervisor (care home), nights

Sunny Meadows Care, 60 residents, 14 robots, 1 of you.

You’ll oversee our overnight fleet of lifting, fetching and cleaning robots, step in whenever one gets stuck (they get stuck), and do the parts of the job no machine is allowed to do: sitting with a resident who can’t sleep, noticing when someone isn’t themselves, and making the call on whether to wake the doctor.

Must have: a care qualification, calm hands, and the patience to reboot a hoist at 3am without swearing in front of Mrs Patel.

The real trend: On this site aged care workers are among the jobs that hold up best, because the work is hands-on and caring. People are building robots to do the lifting and fetching, and whether anyone wants a machine doing the caring part is as big a question as whether one could. This ad guesses at the answer: machines do the carrying, people do the caring.

2. Home energy retrofitter (heat pumps, batteries, the lot)

Retrofit crews, nationwide. Van, tools and tablet provided.

Every house built before 2025 needs taking apart and putting back together more efficiently. You’ll survey old houses, fit heat pumps and home batteries, and explain to owners why their radiators need to be bigger. The AI does the heat-loss calculations. You do everything that involves a loft, a crawl space or an argument about where the unit goes.

Must have: plumbing or electrical background. Must not be scared of spiders.

The real trend: The plumber and electrician profiles already list heat pumps, home batteries and regulation-driven retrofit work under “what’s arriving”. It’s regulation-driven work in cramped, fiddly spaces, which is exactly the kind of work machines find hardest.

Mid-sized firm. Law degree required. Nerves of steel preferred.

Our systems draft contracts, review documents and flag risks faster than any team we’ve ever had. What they can’t do is be accountable. You’ll be the human who reads, questions and signs, and whose name is on it when it matters.

Must have: the ability to say “no, that’s wrong” to a machine that is right 98% of the time, and to know which 2% you’re looking at.

The man in a suit reads a long contract at a polished desk, pen raised to sign, while the robot waits nervously beside him with the printout still curling from its chest.
The robot drafted it. The human's name goes on it.

The real trend: Drafting is among the most exposed parts of legal work on this site, which is part of why paralegals score lower than lawyers. Accountability doesn’t move to software, because you can’t take a model to court. The more drafting the tools do, the more the job becomes the checking and the signing.

4. Apprentice master (AI-assisted trades academy)

Teaching workshop. Part teacher, part tradesperson, part referee.

Our apprentices learn from AI tutors that never get tired of explaining. They still need someone who’s actually done the job to show them how it feels when a joint is right, and to tell them honestly when it isn’t. You’ll run a workshop of twelve apprentices and their tutors, and sign off every one of them.

Must have: ten or more years in a trade, and the patience of a saint.

The man, as a grey-templed master carpenter, runs his thumb along a dovetail joint held up by a young apprentice, while the robot shows a diagram of the same joint on its screen.
The tutor can explain a dovetail. It takes a thumb to tell you it's right.

The real trend: The research on AI at work so far shows it helps beginners much more than experts, and early hiring data suggests junior roles are shrinking in the most exposed jobs. If AI can tutor, experienced people become more valuable as the ones who judge whether the learning worked.

5. Classroom lead, age 7 to 9

Primary school. Classes of 28 humans and one very well-behaved assistant.

The AI marks, plans and differentiates. It knows exactly which child is stuck on fractions. You do what a class of eight-year-olds actually needs: keep order, notice who’s upset, turn a wet Tuesday into something they’ll remember, and make sure nobody eats the glue.

Must have: teaching qualification, loud voice, a sense of humour.

The man, as a teacher, crouches beside two small children at a low desk, pointing at their drawing, while the robot hands out worksheets behind them. A glue pot sits open on the floor.
Somebody still has to keep an eye on the glue.

The real trend: Primary school teachers hold up well here because supervising children and keeping order are barely touched by AI. Planning lessons and marking are the most exposed tasks. Take those away and what’s left is the part most teachers say they came for.

6. Robot referee (warehouse)

Big shed. Lots of robots. Occasional disagreements.

When forty autonomous machines share one warehouse floor, sometimes they all want the same aisle. You’ll watch the fleet, settle traffic jams, spot the robot that’s quietly been doing something daft for an hour, and walk the floor to fix what the cameras can’t see.

Must have: good eyes, steel toecaps, and a whistle.

The real trend: The new jobs page already lists robot teleoperators and data centre technicians: people whose job is to keep machines running rather than to be replaced by them. More machines has so far tended to mean more people looking after machines.

So what’s the pattern?

Read them together and they rhyme. In every ad the machine does the routine, measurable, repeatable part, and the person does the parts that need a body in the room, a judgement call, or someone to be accountable when it goes wrong.

That’s not a prediction; it’s just what the evidence on this site points towards. Which of these would you apply for? And what would the ad for your job say in 2040?

What this article is built on

Every source on the site, with its licence, is on the methodology page · data 0.8-observed-use

The jobs in this article

  1. Not ExposedWritten profile Plumber Almost all of the work is measuring, cutting, joining and fitting pipes on site, which no AI tool can do. Reading the plans is the only task it gets anywhere near. 85 holds up 59 AI helps 15 AI does it 72.1 Hands-on
  2. Minimal ExposureWritten profile Aged care worker Helping clients wash, dress, eat and move AI barely touches. What it could take on is the admin around the care: booking appointments and keeping care records. 75 holds up 47 AI helps 21 AI does it 61.7 Hands-on
  3. Not ExposedWritten profile Electrician Installing, wiring and inspecting on site make up most of the job, and AI barely touches six of the eight tasks. Only reading wiring diagrams and testing circuits come anywhere close, and even those are mostly out of reach. 81 holds up 57 AI helps 16 AI does it 76.1 Hands-on
  4. Exposed: Gradient 1Written profile Paralegal Drafting legal documents and preparing property and share transfers are where AI could take a fair chunk. Keeping order in court and investigating theft it barely touches, which pulls the average back up. 61 holds up 59 AI helps 37 AI does it 14.7 Hands-on
  5. Not ExposedWritten profile Primary school teacher Supervising children and keeping order AI barely touches. Planning lessons and marking are where it gets closest, and even those are mostly out of its reach. 74 holds up 63 AI helps 32 AI does it 23.4 Hands-on

Read next

  1. Work that is forming 4 min read

    Job titles lag behind the work

    The occupation lists this site runs on are stable by design, so they cannot name a job that is two years old. What that hides, and how to read a score for work whose title has not caught up.

  3. Choices 4 min read

    Where do apprentices come from now?

    The tasks AI does best are often the ones juniors learned on, and young workers in exposed jobs are already being hired less. If nobody climbs the first rung, who reaches the top?