Recent social media posts and articles have claimed that Google quietly changed its policies to “greenlight” the use of users’ Gmail messages and attachments for training its AI models, with the only way to opt out being to turn off “smart features” (such as spell-check).

According to a report published today by The Verge, Google has pushed back against these claims and issued a clarification. Google spokesperson Jenny Thomson stated:

“These reports are inaccurate — we have not changed any settings. Gmail smart features have existed for years, and we do not use the content of users’ Gmail messages to train Gemini AI models.”

modezone.com notes that The Verge still recommends users double-check their own settings. Testing showed that even after opting out of certain “smart features,” the system sometimes re-enables them later. In January of this year, Google updated the personalisation settings for innovative features, allowing users to independently disable them for Google Workspace apps, as well as other products such as Maps and Wallet.

Beyond spell-check, enabling Gmail’s innovative features provides several conveniences, including automatic order tracking and adding flight details to your calendar. According to the settings page, turning on Workspace features means “you agree to let Google Workspace use your Workspace content and activity to optimise your experience.” Still, Google has clarified that this does not mean your email content will be used for AI training.

The above content is compiled by ModeZone, a fashion and entertainment magazine.