Skip to content
STIRUP

TECHNOLOGY INTELLIGENCE

Understand what's next.

Gemini Updates Allow Users to Limit AI Training Data Usage

Google has introduced new settings in its Gemini app that let users limit how their data is used for AI training, offering greater control and transparency over personal information.

Share
AI-generated conceptual illustration accompanying: Gemini Updates Allow Users to Limit AI Training Data Usage
AI-generated editorial illustration of the article topic; not a photograph, product screenshot or measured result.

New Controls for Limiting Data Usage in Gemini

Google is enhancing user control over data privacy in its Gemini app by introducing new settings that limit how the AI uses user data for training. Previously, options to restrict training on personal data were limited, but the company is now rolling out changes that separate chat history from training processes. This allows users to maintain basic AI functionality without granting Google extensive access to their data. The adjustable history limit remains available, but users no longer need to modify it to prevent their Gemini usage from being used for training purposes. These updates aim to provide greater transparency and control over personal information, though they do not eliminate all data sharing entirely. The changes are currently being implemented and are expected to improve user confidence in the platform's privacy practices.

Previous Limitations on Opting Out of Data Training

Previously, users who wanted to avoid contributing their data to Google’s AI training had limited options. According to Ars Technica AI, the only way to opt out was to disable chat history, which also removed access to previous conversations. This approach was criticized for being overly restrictive, as it prevented users from benefiting from AI features while still ensuring their data was not used for training. Ars Technica AI notes that any data previously shared with Google’s AI systems is already part of the training data, highlighting the challenge of fully opting out after initial use. Furthermore, Ars Technica AI clarifies that the data collected includes not only user inputs and AI responses but also uploaded files, tool calls, location data, and information from connected apps. These details underscore the breadth of data that can be incorporated into AI training, even when users attempt to limit their contributions.

Data Retention and Usage Beyond Training

Google has introduced new settings that allow users to opt out of having their data used for training AI models like Gemini. However, opting out does not guarantee complete data erasure. According to Google, user data may still be retained and used for other purposes for up to three years. This includes data from interactions with Google services, even if history is disabled. Currently, if users disable history, Google retains that data for 72 hours (and 24 hours for the user), as noted in internal documentation. These policies highlight the distinction between data retention for AI training and broader data usage. Users are advised to review Google’s privacy policies for full details on how their data may be handled beyond AI training purposes.

Limitations of Opting Out of Data Retention

Google has made it easier to use Gemini without training AI on your data, but limitations remain for those who opt out of data retention. According to the company, even if users disable history, Google will still retain some data to "respond to you, process your feedback, and protect Google, its users, and the public." This allows the company to identify potentially harmful AI usage, even when history is disabled. However, chats that have already been reviewed and included in the training data will not be retroactively deleted. This means that past interactions may still contribute to AI training, despite user preferences. As such, opting out of data retention does not fully eliminate the possibility of personal data being used for training purposes. Users are advised to review Google's privacy policies for more details on how their data is handled.

Manual Data Removal and Privacy Options

Google has introduced new manual data removal and privacy options for users of its Gemini AI, allowing greater control over personal data. According to Google, users can manually remove any data they do not wish to be saved on the same page where they manage history and training toggles [Ars Technica AI]. Additionally, users can access these settings through their Google account under Data and Privacy > My Activity > Other activity > Gemini Apps Activity [Ars Technica AI]. Within these settings, users will find a toggle to block Google from using their chats for AI training [Ars Technica AI]. These features provide users with more transparency and control over how their data is used, though they do not eliminate the possibility of data being used in other ways outside of Gemini training.

Sources and reporting notes

Based on published source material. No independent hands-on testing is claimed.

Read more