Two bad answers, and a useful one
If you have been reading about AI and work, you have been handed two stories. One says nothing much will change and you should calm down. The other says everything goes away and you should panic. Neither is a straight answer, and neither gives you anything you could check.
Here is a more useful frame, and it is the one this whole course sits on. AI is not replacing jobs. It is replacing tasks. A job is a bundle of tasks, and the bundles are being reshuffled. Some jobs are made almost entirely of the tasks that shift easily, and those roles genuinely shrink. Some jobs contain almost none, and they change very little. Most people are in the middle, doing work where a slice of the week moves and the rest stays exactly where it was.
In plain English
- Task-level change:
- Looking at the individual activities inside a job rather than at the job title, because activities are what actually move.
- Draft-tolerant work:
- Work where a rough first version is genuinely useful, because a person will edit it before it counts.
- Augmentation:
- The tool does part of the work while a person keeps control of the result. Far more common than outright replacement.
- Accountability:
- Someone whose name is on an outcome and who answers for it when it goes wrong. Software cannot hold this.
The three properties that predict movement
A task tends to shift toward AI when all three of these are true at once.
It is repetitive, meaning you do a version of it again and again and the shape stays the same. It is text-shaped, meaning the input and the output are mostly words, numbers or code rather than physical things or a room full of people. And it is draft-tolerant, meaning a rough first version has real value, because someone is going to read and fix it anyway.
Miss one of those and the task tends to hold. Summarising a long report is all three. Sitting with an upset customer is none of them.
Notice how much of the honest answer depends on your specific week rather than your job title. Two people with the same title can have very different exposure, because they spend their days differently.
What is not going anywhere soon
Some work stays stubbornly human, and it is worth knowing why rather than just being told to relax.
Work that needs someone accountable. If a decision has consequences that someone must answer for, an organisation needs a person attached to it. A tool that produces an answer cannot be held responsible for that answer, and that gap is not a technical problem waiting to be solved.
Work that needs a body in a place. Fixing, installing, caring, building, inspecting, moving things.
Work that runs on relationships. Trust is built slowly between specific people, and it does not transfer.
Negotiation, and judgement someone has to defend. Deciding what should be done when interests compete and the information is incomplete, then standing behind that decision in a room full of people who might disagree.
Honesty cuts both ways. Roles that are heavy in repetitive, text-shaped, draft-tolerant work do get smaller, and pretending otherwise would be a disservice. If that describes most of your week, it is far better to know now, while you still have time and an income. The rest of this course is about what to do with that knowledge.
Work out your own exposure
The useful move is not to read another forecast. It is to audit your actual week. Be careful about asking AI to predict your future, because it will produce confident nonsense on request.
โ Weak prompt
Prompt
Will AI take my job? I work in marketing.
Output
Marketing roles are likely to be significantly transformed. Many routine functions may be automated, though creative and strategic work should remain in demand.
Vague, impossible to check, and roughly true of any job. It gives you nothing to act on.
โ Good prompt
Prompt
Here are the 12 things I actually did at work last week, with rough hours. For each one, tell me whether it is repetitive, text-shaped and draft-tolerant, then sort them from most exposed to least.
Output
Highest exposure: drafting the weekly update (repetitive, text, draft-tolerant, 3 hours). Lowest: the supplier call where you renegotiated terms (relationship, judgement, live).
Specific to your week, checkable against your own memory, and it points straight at what to change.
I am going to list everything I did at work last week, with rough hours for each.
For each task, score it on three things: is it repetitive, is it mostly words or numbers rather than physical or in-person, and would a rough draft still be useful because I would edit it anyway.
Then sort the tasks into three groups: likely to shift toward AI, partly shifting, and unlikely to shift. Explain each placement in one sentence.
Do not predict dates. Do not give me percentages. Do not reassure me. Be blunt.
MY WEEK:
[paste your list]
Once you have that list, run the opposite question. The point is to find where your own reasoning is weak, not to collect comfort.
Here is my task exposure list from earlier. Argue against it. For every task I put in the low exposure group, give me the strongest case that I am wrong and it is more exposed than I think. Then do the reverse for the high exposure group.
[paste your list]
Checkpoint
AI moves tasks, not job titles. The tasks that move are repetitive, text-shaped and draft-tolerant. Work needing accountability, physical presence, relationships or defensible judgement stays put.
The quiet question worth asking
You do not need to raise anything dramatic with your manager. A low-key, practical version works better, and it makes you look thoughtful rather than nervous.
Help me phrase a low-key question for my manager about how our team plans to use AI. I want to sound curious and useful, not anxious, and not like I am fishing for news about job security. Give me three versions with different levels of directness, each under 40 words.
Knowing which parts of your week are exposed is not a comfortable feeling. It is, however, the only starting point that leads anywhere useful. Next we look at the other side of the ledger: the skills that get more valuable, not less, once everyone has these tools.
๐ Quiz
Question 1 of 4What is the more accurate way to think about what AI is changing at work?