Ask it properly · 6 min read
Can ChatGPT help me choose a career?
A general AI assistant is genuinely good at some parts of this and structurally bad at others. Here is which is which, the prompts worth using, and what to do with the answer once you have it.
Written by Nivaan, founder of OffLadder · Last reviewed · Sources listed at the foot of this page
The short answer
ChatGPT and other general AI assistants are useful for the informational part of choosing a career: explaining what a job involves, naming directions you had not heard of, drafting outreach messages, and turning a vague interest into a concrete first task. They are weak at the deciding part, because they cannot see what you actually do, they answer in the confident voice of whatever you asked for, and they have no record of what you tried last month or how it went. Use an assistant to widen the field and design the experiment, then judge the direction by what happened when you ran it.
It is now the first thing most people do, and it is not a silly instinct. An assistant will patiently explain any job in the world at whatever level of detail you ask for, at two in the morning, without making you feel behind. That is a real improvement on having nobody to ask.
What it is genuinely good at
- Explaining the shape of a job: what the day contains, what the words mean, what the entry routes are called.
- Widening the field. Ask for ten adjacent kinds of work that use the same abilities and you will meet titles you had never heard.
- Drafting. The message to a stranger, the application paragraph, the awkward question to a tutor.
- Designing a small experiment, if you ask for one specifically: the one-hour version of the work rather than the reading list.
Where it structurally cannot help
Three limits are worth naming plainly. It agrees with your framing, so it will argue for whatever direction your question implies. It has no evidence about you — only your self-description, which is the same weakness personality tests have. And by default it forgets: the thing you tried and hated last month does not shape what it suggests today, so the advice stays general no matter how many times you ask.
That last one matters more than it sounds. Choosing a direction is not one conversation; it is a sequence of small findings that should narrow the field. Advice that starts fresh every time cannot narrow anything.
Prompts that produce something useful
- 01"Describe a normal Tuesday for a [role]. Include the repetitive and unglamorous parts, not just the interesting ones."
- 02"Give me ten kinds of work that use [thing I am good at] where I would not have thought to look."
- 03"What is the smallest real task inside [role] I could complete in one hour tonight, that produces something?"
- 04"Argue against me going into [direction], given [my situation]. Be specific."
- 05"Who near me might already need this, and what would I say to them?"
Notice that every one of those ends in something checkable. That is the test for whether you are using an assistant well: does the answer give you something to do, or just something to feel.
An assistant can widen the field. Only trying it narrows it.
Questions people ask
- Is ChatGPT good for career advice?
- It is strong on explanation and drafting, and weak on judgement about you specifically. It will tell you what an acoustic consultant does, name adjacent roles you had not considered, and write you a decent email to one. It cannot tell you whether you would enjoy the work, because it has no evidence about you beyond what you typed.
- Why does AI career advice feel generic?
- Because it is answering a general question with a general answer, and because it agrees with the framing you brought. Ask 'should I go into data' and you will get the case for data. The fix is to ask it for the unglamorous 60% of the job, the reasons the direction might not suit you, and the smallest real task you could try this week.
- What should I ask an AI assistant about careers?
- Ask what a normal Tuesday looks like in the role, which parts are repetitive, who actually hires for it and by what route, what adjacent work uses the same abilities, and what one-hour task would reveal whether you like it. Those answers are checkable. 'What career should I do' is not.
- How is OffLadder different from asking ChatGPT?
- It keeps a record and it ends in something to do. OffLadder remembers what you tried, what you liked and what you ruled out, uses that to suggest directions you would not have thought to ask about, and gives you one small real experiment rather than a paragraph of advice. It is deliberately not in the business of telling you who you are.
- Should I trust what an AI says about which jobs are safe from AI?
- Treat confident lists of AI-proof jobs as opinion, whoever publishes them. The honest position is that task mixes are shifting and nobody knows the ten-year picture, which is an argument for testing directions cheaply rather than betting on a prediction.
Directions you could test
Sources
- Future of Jobs Report 2025 · World Economic Forum
- Generative AI and Jobs: a global analysis of potential effects on job quantity and quality · International Labour Organization
How this page is written
We only state what can be checked. Where a page makes a claim about work, automation or the labour market, it links to a primary source rather than paraphrasing a secondary summary. We do not publish predictions dressed as facts, invented statistics, or lists of AI-proof jobs. Pages are reviewed and dated when the argument or the underlying evidence changes. More about OffLadder.