Replace a broad worry with a scenario you can prepare for: what might change, how it could affect you, what evidence would signal a change and what you could realistically do next. A career risk plan is useful even when the probability is unknown.
You do not need to predict the labour market or become “irreplaceable.” Start with the exposure that matters to your current decision, such as reliance on one type of task, an uncertain qualification route or a possible change in working hours.
Separate a signal from a forecast
Write down what is known and where it came from. An official announcement, a change to your actual responsibilities and an opinion about an industry are different kinds of information. A rumour is a question to investigate, not a confirmed event.
For AI, distinguish what a tool might technically do, what your organisation has adopted and how responsibilities are actually changing. ILO cautions that occupational exposure indicators should not be read as predictions of job losses, wages, productivity or retraining needs. They do not tell you that your particular job will disappear on a particular date.
Ask what remains uncertain before deciding on a response. You may need to check a manager's announced plan, a current local qualification rule or the actual requirements in target vacancies. Avoid repeatedly reading general forecasts when a specific question has a closer source.
Build one scenario at a time
| Scenario | My exposure | Evidence and unknowns | Signal to watch | Proportionate first action | Fallback and review |
|---|---|---|---|---|---|
| A routine task changes | Much of my work uses the same process | Tool trial announced; future role not decided | Confirmed change to task ownership | Learn the new checks within authorised work |
These are illustrative scenarios. Fill in your own evidence rather than assigning invented numerical probabilities. Consider impact, time to respond and what you can change; a risk outside your control may need a fallback rather than another training course.
Worked example: prepare for task change without assuming redundancy
This is a fictional teaching example. Priya, the organisation and observations are invented. There was no employer study, forecast or measured intervention result.
Priya spends much of her week assembling a recurring report. Her employer announces a trial of a tool that may automate parts of the process. There is no announcement about staffing or final task ownership. Priya's first interpretation is “my job is ending.” Her plan begins by separating that fear from the actual announcement.
She records three unknowns: which steps the tool covers, who will check its output and whether users still need explanations of unusual results. She asks her manager about the trial rather than putting company data into an outside service herself.
With permission, she lists her current steps: collect inputs, check missing entries, assemble the output and explain exceptions. In a small authorised test, the draft output omits an exception that the current manual check catches. The invented observation shows a quality issue in that test; it does not establish the tool's general capability or guarantee that her checking work will remain a job.
Her immediate action is to document the exception rule and ask who will own it during the trial. She keeps the approved existing process until a replacement is accepted. Separately, she prepares a non-confidential description of her actual contribution and investigates target roles that use similar checking and explanation tasks.
She sets a review for the announced trial decision. If responsibilities change, she will compare the new expectations with her skills and available development support. If there is no decision, she updates the uncertainty instead of treating delay as proof that the risk has vanished. If a formal employment change is announced, she will check the actual terms and appropriate local advice rather than rely on this worksheet.
The result is a small preparation step and a clearer trigger. It is not a promise of job security, a prediction of dismissal or a reason to resign pre-emptively.
Choose a response that does not create a larger exposure
Match the action to the uncertainty. If the issue is missing role information, obtain that information. If it is a task-specific skill gap, choose bounded practice and feedback. If an entry rule is decisive, check the authority before buying training.
Consider your time, income obligations, caring responsibilities and health. A second income stream can add costs, restrictions and workload; it is not automatically a low-risk solution. Check relevant employer and local rules before outside work. Do not count hoped-for freelance revenue or an unconfirmed offer as secured income.
Use your own financial circumstances when considering a buffer. This guide does not prescribe a universal number of months or recommend investments. Significant financial, legal or health decisions may need qualified advice that applies to you.
Give the plan a trigger and a limit
Write “If [observable event], I will [feasible action] by [date].” Also write what would make you stop, reduce or replace the action. A course that exceeds your budget or does not address the target requirement should not continue simply because you already started it.
Keep one short record of the source, action, person or authority needed, review date and fallback. Reopen it when relevant information changes; do not turn uncertainty into a daily scoring exercise.
Your next action is to complete one scenario row and verify its most important unknown. The workplace skill guide can help turn a confirmed gap into practice. The annual review guide helps compare existing evidence, and the career planning hub connects the next step with broader options.
Source and scope
ILO: Workers’ exposure to AI—what indicators tell us and what they do not, published 17 April 2026, read 22 October 2026. Used for the limits of AI exposure indicators, not to predict any reader's employment outcome. The scenario table, triggers and fictional example are editorial planning tools, not a validated risk score or personalised employment, financial or legal advice.