AI job risk checklist: track evidence before changing careers
An AI job-risk checklist should help you record what is changing and what remains uncertain. It should not output a made-up probability of losing your job. Use observable workplace changes, realistic alternatives and a review date to decide which action deserves attention.
Record the signal and its reliability
Write the change, where the information came from and what it actually implies. A new tool trial, a reduced task and a formal restructuring announcement are different signals. Avoid treating a social-media prediction as evidence about your employer’s plans.
Check your current evidence
List the outcomes you help deliver and the judgement needed to deliver them. Identify which contributions remain useful if drafting or routine analysis becomes faster. Note whether you can show that work through an approved example, a documented result or feedback from colleagues.
Compare actions before choosing a large move
Illustrative options: learn an approved evaluation workflow, seek responsibilities involving exceptions, investigate an adjacent function or begin a focused external search. Record what each option costs in effort and what uncertainty it addresses. Buying a generic AI course may answer none of those questions.
Set a review trigger
Review when there is meaningful new evidence: a changed responsibility, feedback on a sample or clearer requirements for an alternative role. Keep separate notes for observations and assumptions. Use the task-audit article if you still need to understand which parts of your work are exposed.
Create a signal-to-action record
Use five columns: the observed signal, its source, the implication you think it has, the missing evidence and a proportionate action. For example, a tool pilot may justify learning how its outputs are evaluated. It does not automatically justify assuming a restructuring has been decided.
Review the record when new information arrives. Remove assumptions that were not supported, update confirmed changes and identify whether an alternative role needs investigation. Keep the action tied to the evidence rather than to an alarming prediction. This record helps you prepare while preserving uncertainty. It is not a model that calculates individual employment risk or tells you when to resign.
Choose the next guide for your question
Sources and further reading
Editorial review: 13 October 2026. Examples are illustrative; results depend on the information supplied.