Gemini and AI content

Can a human read your Gemini chats?

Yes, and Google says so in plain words:

• Trained reviewers, including people at service providers, read and annotate a sample of conversations.
• Google’s own warning: "Please don’t enter confidential information that you wouldn’t want a reviewer to see or Google to use to improve our services, including machine-learning technologies."
• The setting that controls this is now called Keep Activity, renamed from Gemini Apps Activity.
• Turn Keep Activity off and chats are still held about 72 hours to run the service. Leave it on and activity auto-deletes at 18 months by default, adjustable to 3 or 36 months, or never.
• Reviewed chats are the catch: once a conversation is sampled for human review, it is disconnected from your account and kept separately for up to three years, and deleting your chat does not pull it back.
• Temporary chats are the exception: Google says they are not used to train its AI models or improve its AI with help from human reviewers.

Does Google watermark AI content from Gemini?

Google applies SynthID, an invisible watermark embedded in the content itself, to media from its generative models. It is designed to survive some edits, and Google publishes a detector for it. Google’s generative AI terms cover what you may do with the output.