Think you might be an AI ethicist?
A lot of people doing this work already don't call themselves this yet. Nine honest questions, loosely mapped to the knowledge, skill, and attitude the role actually requires. No email required, and there's no failing grade.
Where you are today
Answer how things actually stand, not how you'd like them to read on a CV.
1. When a system’s output could disadvantage a particular group, can you say specifically which group and why, not just that “bias” is a risk in general?
2. If softening a finding would keep the team you’re reviewing comfortable, and nothing on paper requires you to choose, does your obligation run to the people affected by the system, or to the relationship with the team?
3. Have you had a flagged feeling about a system that you couldn’t yet justify in a slide, and followed it anyway instead of waiting until it could be footnoted?
4. Have you had to explain an AI-driven decision to someone affected by it, in terms they could actually use, not a compliance-style notice?
5. Have you walked into a room to deliver a finding you knew the people in it did not want to hear, knowing it could cost you the relationship or the contract, and said it anyway?
6. Have you pushed a system to do more good, not just less harm, improving access, fairness, or an outcome, rather than only avoiding a headline risk?
7. Have you had to build in, or argue for, a point where a system hands the decision back to a person rather than carrying it through alone?
8. Have you told someone that “it’s legal” isn’t the same as “it’s fine,” and held that line even though nothing on paper required you to?
9. Have you disagreed with someone reasonable about the right call on a case, been able to see genuine weight in their position, and still made a decision anyway?
10. Have you caught a system designed to feel more trustworthy, empathetic, or available than its actual reliability warrants, and flagged what that design does to the people relying on it, not just whether its outputs were biased?
A rough gauge against the knowledge, skills, and attitudes the role actually requires, in the same independence-first model we place people against, not a pass/fail exam and not a full competence assessment.
It's a rough mirror, so you can see which of knowledge, applied skill, impartiality, attitudes, autonomy, positive impact, bravery, judgement, and pluralism is already there and which still needs building, before you put yourself forward. Not a certificate, and it doesn't feed anywhere, just a clearer picture of where you stand.
However that landed, there's a route in.
Pupillage exists for exactly this: real hours, under an accredited mentor, building the evidence as you go.