We are building an application for a pharma client using Gemini APIs, and data privacy is a critical concern for us due to the highly confidential nature of the data.
From the documentation, we understand that data sent via Gemini APIs is not used for model training by default. We would like to confirm a few points:
-
Is it correct that data shared through Gemini APIs is not used for training the models?
-
Does Google store request/response data when using the Gemini API?
-
If yes, what is the retention duration and purpose of storing this data?
-
For sensitive healthcare/pharma use cases, is using Vertex AI with zero data retention the recommended approach?
We want to ensure full compliance and follow best practices before deploying our solution.
Any clarification or official guidance would be greatly appreciated.
Thank you.