Does NotebookLM train on your data?

No. Google has explicitly stated that NotebookLM does not use your uploaded sources to train its AI models. Your documents, notebooks, and conversations are not used as training data for Gemini or any other Google model.

NotebookLM processes your sources in context — they are available to the model during your session to generate responses, but that processing does not feed back into model training. Your data is also not visible to other users.

Google does collect product usage data (how you interact with the interface) as described in its standard privacy policy, but the content of your sources stays private to your account. For sensitive documents, this makes NotebookLM safer than tools that do use conversation history for training.