Google has rejected widespread claims suggesting that Gmail messages and attachments are being used to train its Gemini artificial intelligence model. The company responded after several posts and articles circulating on social media, including a report by Malwarebytes, alleged that Google had updated its policies and that users could only opt out by turning off Gmail’s smart features.
IMPORTANT message for everyone using Gmail.
— Dave Jones (@eevblog) November 19, 2025
You have been automatically OPTED IN to allow Gmail to access all your private messages & attachments to train AI models.
You have to manually turn off Smart Features in the Setting menu in TWO locations.
Retweet so every is aware. pic.twitter.com/54FKcr4jO2
Google clarifies its position
A Google spokesperson, Jenny Thomson, said that the circulating information was inaccurate. She said, “These reports are misleading – we have not changed anyone’s settings, Gmail Smart Features have existed for many years, and we do not use your Gmail content for training our Gemini AI model.”
The company emphasised that it has not made any changes that would automatically restore users’ settings. However, one staff member reported that, despite previously opting out of certain smart features, they later found those settings had been re-enabled. Google has not confirmed any technical issue but noted that users have complete control over their personalisation options.
Google introduced an update to its smart feature personalisation settings in January, allowing users to manage settings for Google Workspace separately from other Google services such as Maps and Wallet. This split was intended to give customers clearer control over how their data is used across Google’s ecosystem.
How Gmail’s smart features work
Gmail’s smart features include tools such as spell checking, automatic order tracking and functions that help users add flight details directly from an email into their calendar. When enabled, these tools rely on data within a user’s Google Workspace account to personalise their experience.
The settings page for Workspace explains that turning on these features means “you agree to let Google Workspace use your Workspace content and activity to personalise your experience across Workspace.” Google has clarified that this personalisation is separate from using Gmail content to train AI models. According to the company, the smart features operate locally within the Workspace environment and do not feed into Gemini’s training datasets.
These features have been part of Gmail for several years and are designed to streamline everyday tasks. Google maintains that its use does not involve contributing email content to AI training efforts, despite online speculation suggesting otherwise.
Users encouraged to review settings
Although Google denies the claims, the company suggests that users regularly review their personalisation settings to ensure they reflect their preferences. By navigating the Workspace and Google account settings, users can independently adjust personalisation options for email, maps, payments and other Google tools.
The company reiterated its commitment to transparency regarding how data is used across its services. It also stated that Gmail users retain control over whether smart features are active and how much personalisation they allow within Google Workspace.
Google’s response aims to counter misinformation and reassure users that their private Gmail content is not being accessed to develop or train the Gemini AI model.
