How to tell if AI will replace your job (and what to do about it)
The question "will AI replace my job?" is the wrong question. The better one is: which of my tasks is AI automating, which is it augmenting, and which stays mine? That task-level view tells you exactly what to do next.
Think in tasks, not job titles
Job titles are too coarse to predict AI impact. A bookkeeper and an accountant sound similar, but a bookkeeper spends far more time on automatable transaction categorization, while an accountant spends more on advisory and judgment.
Break your job into 6–15 weekly tasks and ask of each: could a capable AI do this today, or within two years, without me? If yes, it is automatable. If AI speeds it up but you still lead, it is augmentable. If it needs relationship, physical presence, accountability, or novel judgment, it is human-essential.
Weight by how you spend your week
A task being automatable does not matter much if it is 2% of your week. It matters a lot if it is 30%. The share of your week in each category is a better signal than any single task.
If more than half your week is automatable, you have an urgent reskilling case. If most is augmentable or human-essential, your job is reshaping, not disappearing — and AI may make you more productive.
Build the skills AI cannot do
The most future-proof skills combine judgment, relationships, and AI fluency. Judgment means deciding under uncertainty. Relationships mean trust, empathy, and negotiation. AI fluency means directing AI effectively and evaluating its output.
Pick one skill from each category and build over 90 days. You do not need to learn to code — you need to learn to direct AI, judge its output, and do the human work it cannot.
Start with a 90-day plan
A 90-day plan beats a vague resolution. Spend the first four weeks mapping your exposure and building AI fluency. The next four, apply AI to real work and ship something. The final four, teach a teammate and propose one improvement.
The point is not to become an AI expert. It is to move from passive anxiety to active competence.
FAQ
Estimates based on task-level AI capability research, not a guarantee of employment outcomes.