"Generated Data, output of audio recordings to generate Summaries, transcripts and texts, created through AI models. Translations of your Summaries or Transcripts, Speakers names. Voice Generation from summaries. Metadata related to AI interactions (e.g., length of calls, length of transcription)."
User audio recordings, meetings, and calls are processed through AI models to generate transcripts and summaries. While framed as service delivery, the policy does not explicitly rule out using this data to train or improve AI models.