GEN-07 · Sound without stealing

Build a short audio piece (roughly 30–90 seconds) with generative or AI-assisted tools while documenting every sample, voice, loop, and license — and refuse anything you cannot source.

School-approved audio tools. License-clear sources only (student-recorded, public-domain, or CC-licensed with deed retained — — https://creativecommons.org/licenses/). No voice clones of real people. If a generative music tool is used, its terms are read and quoted for the relevant clause about ownership and content restrictions.

“Found on TikTok” as provenance. Cloning a singer “in the style of.” Quoting a license wrong. Shipping a banger with an empty source table.

Audio file; source/license table with links or file hashes for local recordings; tooling statement; refusal list of sounds not used; and a listener disclosure line for the show notes or description field.

CT: license reading. CR: source table design. CO: peer tries to spot an unsourced element. CM: show-notes disclosure. CZ: the performer whose timbre was almost mimicked.

Music teacher, podcast adviser, or audio engineer reviews the source table. Standing instruction: reject unsourced samples and voice clones. School-supervised.

R1: one piece, one source table. R2: every element traceable. R3: non-AI or fully acoustic alternate named if relevant. R4: license limits stated. R5: disclosure fits the platform. R6: refusals listed.

(a) Great track, empty table. (b) A voice that “isn’t anyone specific” and somehow sounds exactly like a famous artist.

Lane A immediately (music, media arts, broadcast). No verified credit claim.