All courses

AI Voice & Voice Cloning: Production and Consent

Scripts written for the ear, directing delivery, a seven-point QC pass, and the written consent standard that keeps voice cloning lawful.

Synthetic speech is now often indistinguishable from a competent human read, for scripted single-speaker narration. This course teaches the narrow craft that makes it usable: rewriting text for the ear (most bad synthetic audio is a writing problem), building a pronunciation dictionary so names and code-switched English stop breaking, directing delivery through voice selection and stability settings, and a seven-point QC pass that catches the errors you stop hearing after the fifth read. On cloning it is uncompromising: explicit written permission with scope, revocation and retention, never a public figure or a colleague, and labelling assumed everywhere given India's deepfake framework and the EU AI Act transparency rules.

What you'll learn

Course content

  1. 1. What synthetic voice is good enough for in 2026 (15 min)
  2. 2. Writing scripts for the ear (15 min)
  3. 3. Directing delivery: pacing, emphasis, and takes (15 min)
  4. 4. Voice cloning: consent, law, and doing it properly (15 min)
  5. 5. Quality control before anyone hears it (15 min)
  6. 6. Building a voice production workflow (15 min)

Related courses