AI must not be used for the generation of non-trivial audio content that is directly incorporated into the final work. All creative expression, narration, and performance must remain the original contribution of the human creators.
Specifically, your use of AI must be limited to de minimis levels. That is, uses so minor or mechanical that they do not affect the copyrightable authorship of your work and, in practical terms, do not alter its human-created nature.
Acceptable De Minimis AI Use
For your guidance, examples of acceptable de minimis AI use typically include:
Noise reduction, level balancing, and format conversion
Generating ideas for pacing, delivery, or narration approach
Checking text pronunciation or preparing pronunciation guides
File organization, sample rate adjustments, or metadata tagging
Technical tasks such as equalization (EQ) assistance or removing background clicks and pops
Unacceptable AI Use (Beyond De Minimis)
By contrast, unacceptable AI use beyond de minimis would include:
AI-generated narration or voice synthesis
AI voice cloning or text-to-speech used in the final recording
AI-generated audio content that meaningfully contributes to the final work
AI-generated vocal performance that would be copyrightable if created by a human
AI-generated promotional materials or cover imagery (unless separately disclosed)