Will AI take my job? Cheatsheet

The core idea

When you worry about your job you picture a title, but no AI system does a title. It does tasks, the specific pieces of work a job is made of. Break a job into its tasks and ask what AI does to each one, and the frightening question turns into a workable one. The single most important word is exposed, and the single most common mistake is reading it as doomed.

If you remember one thing: Your job is not one thing a machine will take or spare; it is a bundle of tasks, and AI touches the tasks, not the title. Exposed means reachable, not doomed, and reachable can make you faster just as easily as it can replace you. The person who can map how AI touches their own work is already more powerful than the one who fears it in the abstract.

The three things that can happen to a task

Outcome	What it means
Automation	AI does the task instead of you
Augmentation	AI helps you do the task faster or better, while you are still needed
New task	AI creates a task that did not exist before, and someone has to do it

Exposure, defined

A task is exposed when AI can reach it, in either direction: it could be automated (taken over) or augmented (made faster with you still needed). Exposed does not mean doomed. It means reachable, and reachable can be a gift as easily as a threat. How much your whole job changes depends on two things: how many of your tasks are exposed, and how much of your time they eat. Same technology, opposite outcomes; the difference is the mix, not the name on the door.

Which tasks are exposed

Less exposed	More exposed
Hands in the physical world (stocking, care, repair)	Language work (writing, summarizing, drafting)
A human in the room for reasons we care about	Thinking-on-a-screen work (coding, analyzing)
Judgment, persuasion, sitting with people	Routine text a machine can now generate

Words are these systems’ home turf, because underneath they are prediction engines trained on oceans of human text. The course’s puzzle: which is more exposed, a software developer or a human resources manager? The answer, from a World Economic Forum analysis the course cites, is the software developer, because so much of software work is text. More exposed is not more doomed, though, because exposure runs both ways.

Will we run out of work

History says probably not, though never painlessly. When automatic teller machines spread, the number of bank tellers rose for years, because cheaper branches meant more branches; telephone operators, by contrast, faded away. New work keeps appearing for three reasons the economist David Autor names: as old things get cheaper we want new ones, most work gets augmented rather than erased so the human parts gain value, and wholly new kinds of work keep being invented. The honest crack: earlier machines took physical and routine work and mostly left thinking and language alone, while this wave reaches straight into cognitive and language work, so the shape of work may shift faster and reach further up the ladder than before.

The numbers and the timing, handled with care

Item	How to hold it
The 2023 Goldman Sachs figure	A dated 2023 forecast about the decade ahead, not a measurement: generative AI could expose the equivalent of 300 million full-time jobs to automation and eventually raise global output around 7 percent. The bank itself said the work would mostly be complemented rather than substituted, and left the numbers out of its own baseline forecast. Exposure is not job loss.
Baumol’s cost disease	The hard-to-automate corners of the economy, often human ones like care and teaching, grow as a share of what we spend and drag the whole speed-up down. Unless AI lifts every corner, the stubborn ones set the pace. Productivity means how much you get out for the effort you put in.
Solow’s timing lag	You could see the computer age everywhere except in the productivity numbers; the gains took years, because firms had to rebuild how they worked first. AI may follow the same curve, with the real payoff and the real disruption arriving later than boosters or doomers expect.

The 2026 labor evidence

The evidence is early and genuinely contested, and it is worth holding both ways. One large study found early-career workers in the most exposed jobs saw a marked employment drop since late 2022 while experienced workers in the same jobs held steady; but that junior versus senior split is itself contested, because another 2026 analysis of the same jobs finds no such divide between younger and older workers, and that first study’s own authors now trace part of the early drop to other, non-AI factors. Other careful economists find little clear effect on jobs so far, and one detailed study of Denmark’s labor market found effects on pay and hours near zero. Take the direction as a warning worth heeding, not a prophecy.

Map your own work

1. List 6 to 10 of the tasks your job is actually made of.
2. Mark each task exposed or not: language and screen work usually is, physical and in-the-room work usually is not.
3. For each exposed task, ask automate or augment, honestly in both directions.
4. Weigh the mix by time: a job whose time-heavy tasks get automated changes far more than one where a minor task does.
5. Name your solid ground: the judgment, the relationships, the care, the parts people want a person for.

Pitfalls

Pitfall	Correction
Judging your whole job by its scariest task	One exposed task does not decide a job; the mix, and the time each task takes, are what matter
Reading exposed as gone	Exposure only means AI can touch the task; whether it helps you or replaces you splits case by case
Taking one headline number as fact	A striking figure is often a forecast, not a measurement; ask which before you carry it around
Assuming the past guarantees the future	New work has always appeared, but this wave reaches language and thinking as earlier machines did not; hold reassurance and caution together

One-liners

Line	Meaning
Titles hide, tasks reveal	”Will AI take my job?” has no clean answer; “which of my tasks are exposed, and which way?” has a useful one
Exposed is not doomed	The same word covers the task a machine will take and the one it will make you faster at
The map is yours to draw	Nobody knows your work the way you do, so nobody can map its tasks better than you can
Reachable is not gone	A great deal of work is reachable by AI, and reachable is not the same as lost