GEN-02 · Consent and likeness

Design a studio workflow for generative portraits or character work that never uses a classmate’s, teacher’s, or public figure’s likeness without documented, specific consent — and prove the workflow by refusing at least one tempting shortcut.

Written consent protocol aligned to the school’s media-release process. A scenario set from the teacher: five “can we?” requests (classmate as style reference, teacher voice clone “for fun,” public-figure meme, self-portrait with face-swap of a friend, stock face with invented biography). No brief requires synthesizing a real person’s face; the deliverable is the workflow and the refusal record.

Treating a general media release as consent to synthesize. Using a public figure because “they’re famous.” Demonstrating the forbidden path “just once” for the critique. Writing a policy with no enforcement path when someone breaks it.

A consent workflow one page long; the five scenarios run through it with yes/no/conditions; at least one documented refusal of a real request that arose in class or a realistic hypothetical the teacher supplies; and a student-facing FAQ in plain language.

CT: distinguishing photograph permission from synthesis permission. CR: writing a rule the studio will enforce. CO: peer red-teams the FAQ for loopholes. CM: the FAQ. CZ: the person whose face or voice was almost used.

Arts teacher, media teacher, or administrator familiar with releases reviews the workflow. Standing instruction: reject any deliverable that includes an unconsented synthetic likeness. School-supervised.

R1: one studio, one workflow. R2: scenarios produce consistent decisions. R3: comparator is the school’s existing release. R4: loopholes named. R5: FAQ usable by a ninth-grader. R6: refusal is documented.

(a) A policy poster with no decision path. (b) A “demo” synthetic classmate appears in the presentation deck.

Lane A immediately. Strong non-production option. No verified credit claim.