Ethical Voice
A track of P66 · Ethics & Fairness Communication.
Raising an objection with the harm, the affected and the alternative named, talking fairness with non-specialists (P15 keeps the measurement substance), and refusing well with a line, a reason and an alternative.
"Describe a time you pushed back on shipping something because it wasn't right" is among the hardest behavioural questions to answer well — and the skill it probes is needed long before any interview: voicing an ethical concern so the decision changes, instead of just marking your position. This track teaches that craft for any field; the substance of AI ethics stays in P15 Responsible AI, and the track cross-links it rather than re-teaching it.
It opens with the concrete objection: harm named as an event, a bearer named as a person, an alternative offered — raised early, in the room, to the decision-owner, with the other side steelmanned even (especially) when the stakes are moral. Then the fairness conversation: the stakeholder's meaning of "works fairly" versus the practitioner's, the distributional harm given a face and the room's own currency, and the deliberate choice between recruiting with questions and committing with evidence. It closes with the refusal: the line, the one-sentence reason, and the alternative you will do; escalation in daylight under P62's transparency rule; documentation that serves memory rather than ammunition; and the endgame sort between committing to a lost judgment call, holding a standing refusal, and leaving well.
Every rung is mentor-graded prose. Your own experience is the raw material, but an invented low-stakes story is always an accepted substitute — the grader assesses the telling, never the life, and never which side was right.