“Prohibited” Content

Spoken uses Google's Vertex LLM for analyzing content, and while Spoken has a "whitelisting" agreement with Google to bypass Vertex / Gemini's usual content safety filtering, any content in which it detects child sexual abuse material (CSAM) will not be allowed through their process (or, hence, our analysis process). If you see a Prohibited Content warning on one of your installments / chapters, you will need to alter the content to avoid CSAM filter for that installment in order to generate the narration. If you feel your work has been erroneously flagged, you can get more information on our Ask Patrick help page, or send us an email to look at the project.

Here is how Google describes their process:

https://blog.google/technology/safety-security/how-we-detect-remove-and-report-child-sexual-abuse-material/