Does Google NotebookLM train on your data?

No — Google NotebookLM does not use your content to train its AI models by default. Google states the content in NotebookLM will not be used to directly train its foundational AI models unless you choose to provide feedback. Submitting thumbs up/down feedback routes that content to specially trained human reviewers; eligible Workspace/Education/Enterprise accounts are excluded from human review and training entirely.

There is no training opt-out to set for Google NotebookLM: No training opt-out is required by default because uploaded sources, queries and outputs are not used to train foundational models. The only path that feeds training is voluntarily clicking thumbs up/down feedback, so users simply avoid submitting feedback; Google says do not include confidential or sensitive information in feedback.

For qualifying Google Workspace and Workspace for Education accounts, NotebookLM is a core service under the organization's Workspace agreement and Cloud Data Processing Addendum: uploads, queries and responses are not human-reviewed even with feedback and are not used to train AI models. For enterprise, Google contractually commits under the Google Cloud Terms that documents and outputs are never used to train foundational models.

It depends on your plan tier. Consumer (personal Google account): Sources, outputs and chat history are not used to directly train foundational models; the only exception is content you voluntarily submit via thumbs up/down feedback, which may be human-reviewed (disconnected from your account). Workspace / Education / Enterprise: Covered by the Workspace agreement and Cloud DPA: uploads, queries and responses are never human-reviewed (even with feedback) and never used to train AI models. Always confirm current terms with Google before sending confidential data — this is cited public information, not legal advice.