P66 · Ethics & Fairness Communication

Everyone in the meeting agreed the claim was misleading, and it shipped anyway — not because anyone chose harm, but because the one objection raised was "this is just wrong", and a room can survive an accusation far more easily than it can answer one. Ethical voice is the craft this pillar teaches: not knowing right from wrong, which most professionals manage, but saying it so the decision can change — a skill identical for an interpreter refusing to omit, a teacher questioning a placement test, and an engineer flagging a fairness gap.

The ethical-voice track builds it in three escalating situations. First, the objection: replacing the abstract moral appeal with three named things — the harm as an event, the person who bears it, the alternative you propose — raised early, in the room, to the decision-owner, with the other side's real case steelmanned. Second, the fairness conversation: what "works fairly" means to a stakeholder versus a practitioner, translating a distributional harm into stakes the room already owns, and choosing between questions and evidence; the AI substance — where bias enters, what the conflicting metrics measure — lives in P15 Responsible AI, and this track points there rather than re-teaching it. Third, the refusal: the line, the reason, and the alternative you will do; escalating in daylight; documenting for memory, not ammunition; and the honest sort — commit, carry, or leave — when the call goes the other way.

Throughout, the raw material is your own working life, and feedback assesses the telling, never the life: a rubric may flag "harm not named", never "you were wrong to refuse". An invented low-stakes story is always an accepted substitute — ethics stories can be sensitive, and disclosure is never required.