Announced 4 Nov 2026 · Sources checked
What did Google confirm about Gemini training data?
Google's Gemini Apps Privacy Hub, last updated October 26, 2026, states that Gemini Apps activity is used to improve Google services for everyone, including by training generative AI models, if the Keep Activity setting is on. The setting defaults to on for adult users.
The policy applies to personal accounts regardless of subscription tier. Paid Gemini Advanced subscribers face the same data-use terms as free users. Google's commercial offerings — including the Gemini API, Vertex AI, and Workspace tiers — operate under separate data agreements that exempt customer data from training.
How does the Keep Activity setting work?
Keep Activity controls both whether chats are used for training and whether the assistant remembers past conversations. Turning it off stops future chats from being used to improve Google AI, but the same toggle also deletes stored conversations after 72 hours.
That 72-hour window eliminates long-term memory for the assistant, effectively neutralizing one of the product's core features. Users who want persistent memory must accept that their conversations will be fed into training data.
The all-or-nothing design stands in contrast to some competitors. OpenAI, Google, Microsoft and Anthropic recently agreed on a shared robots.txt signal to let sites block AI training, but the consumer opt-in question for training on user data remains fragmented across providers.
Are human reviewers reading Gemini chats?
Yes. Google's privacy hub confirms that a subset of chats is reviewed by human reviewers, including Google's trained service providers, to help improve services. Reviewers assess whether responses were low-quality, inaccurate or harmful and suggest better ones.
Reviewed chats are retained for up to three years, disconnected from the user's account. Google warns users: 'Please don't enter confidential information that you wouldn't want a reviewer to see.' Deleting activity does not remove chats that have already been reviewed.
Does paying for Gemini change the data terms?
Based on Google's published documentation, a personal paid subscription does not exempt users from training. The Keep Activity setting controls data use the same way for free and paid personal accounts.
This distinction matters because users who pay for a service often assume their data receives different treatment. The Hacker News thread that surfaced the policy on November 4 drew hundreds of upvotes from users who said they had assumed a paid plan meant they were not 'the product.'
Google has faced related questions about consumer data handling before. Its Dreambeans launch in India included a privacy-focused question feature, and the company recently switched Gemini to opt-in for schools after backlash over consumer-style data defaults in educational settings.
How do other AI providers compare?
Google is not alone in using consumer chat data for training. OpenAI also uses ChatGPT conversations to improve its models, though it offers a separate opt-out toggle that does not require disabling chat history. Anthropic's Claude does not train on consumer conversations by default.
The key difference is the binary nature of Google's control: users cannot preserve chat memory while opting out of training. Critics argue this design pushes users toward accepting data use as the cost of a functional product experience.
What can users do to limit training on their Gemini chats?
Users who want to stop training without losing all memory have limited options:
- Turn off Keep Activity in Gemini settings. This stops training but deletes chat history after 72 hours.
- Use temporary chats for individual sensitive conversations. Temporary chats are not used for training.
- Delete past activity manually. Note that already-reviewed chats are retained for up to three years.
- Use Gemini through a Workspace or business account if available, which follows different data agreements.
What does this mean for businesses and developers?
For anyone building with or around Gemini, the practical implication is clear: do not paste customer data, credentials, source code under NDA, or unreleased research into a personal Gemini account with Keep Activity on.
The API and managed Workspace plans operate under separate terms that exempt customer data from training. Developers should verify which plan applies before using Gemini with sensitive information.
As AI products mature, the pressure to offer granular data controls will likely intensify, especially in markets with strict privacy regulation. Google's current binary toggle may face increasing scrutiny from both users and regulators.
Common questions
Can I stop Gemini from training on my chats without losing memory?
Not with the current settings. The only way to stop training is to turn off Keep Activity, which also deletes conversations after 72 hours. Temporary chats offer a per-conversation workaround.
Does Gemini Advanced protect my data differently?
No. Google's published privacy hub does not distinguish between free and paid personal accounts. The same Keep Activity setting and data-use terms apply regardless of subscription tier.
Are my Gemini chats reviewed by humans?
A subset of chats is reviewed by human reviewers including trained service providers. Reviewed chats are retained for up to three years, even after you delete your activity.
Sources & further reading
How this story was made
Written by Kristian Kostov with AI assistance and checked against the linked sources. Company performance claims are attributed to the company. Analysis reflects AiLookout’s interpretation; we have not independently tested the products discussed. The cover is an AI-generated editorial illustration, not a photograph or evidence of an actual event.
Our editorial standards





