Career-specific AI guide
No replacement score. Separate the career’s tasks to see where to direct AI and where to preserve independent reasoning.
Task Automation Exposure
Lower never means permanently safe; higher never means the career is useless. Task mix and responsibility change.
Models can draft the very code and prompts used to call them, but generated patterns are often untested or provider-specific.
Your response: Keep small baselines, typed contracts and regression tests around every generated change.
AI can propose cases and judges, but model-generated evaluation can share the same blind spots as the system.
Your response: Use human-labelled representative cases, deterministic checks and explicit uncertainty.
Someone must decide when the feature should refuse, escalate or not be built.
Your response: Practice threat modelling, privacy review, staged rollout and rollback decisions.
Independent capability
Verification drill
In an AI-generated solution, identify one bug, one unproven assumption, one missing edge case and one simpler alternative.
Next choices