Google Denies Using Your Emails for AI Training: The Truth
▼ Summary
– Google denies allegations that it analyzes private Gmail content to train its AI models, stating it has not changed user settings and has long-standing smart features.
– A class action lawsuit accuses Google of privacy violations for allegedly enabling AI tracking of private communications in Gmail, Chat, and Meet without user consent.
– Malwarebytes initially claimed Google was using email data for AI training but later clarified this was a misunderstanding, noting smart features scan emails for functions like spam filtering and suggestions.
– Three smart feature settings in Gmail are automatically enabled, allowing Google to use content for personalization across products like Workspace, Maps, and Gemini, though users can opt out.
– Users can disable these smart features in Gmail settings on desktop or mobile, but doing so may reduce functionality like Smart Compose and Smart Reply, presenting a trade-off between convenience and privacy.
Google has firmly rejected claims that it scans private Gmail messages to train its artificial intelligence systems, addressing concerns raised by a recent class action lawsuit and security reports. The company insists that its longstanding smart features are separate from AI model training and that no recent policy changes have authorized such data usage. This clarification comes after initial confusion led some observers to believe Google had quietly altered user settings to feed email content into its Gemini AI.
A security company initially reported that Google appeared to be activating features allowing it to review private emails and attachments for AI training purposes. The features in question, Smart Compose, Smart Reply, and predictive text, have been part of Gmail for years. However, revised interface language and placement made it seem like Google might be leveraging user data for new AI training without clear consent.
Google’s official response was unequivocal. A spokesperson stated, “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. Lastly, we are always transparent and clear if we make changes to our terms of service and policies.”
Following this, the security firm updated its original post, acknowledging it had contributed to a “perfect storm of misunderstanding.” The firm clarified that while Gmail does scan email content to power features like spam filtering and writing suggestions, this is standard practice and distinct from utilizing data for AI training. They noted that the settings themselves are not new, but recent wording and interface adjustments led many to believe users were being automatically enrolled in AI training programs.
Despite these clarifications, questions remain about why these smart features are enabled by default. Reports indicate that even users who previously disabled certain options found them reactivated without their knowledge. Upon inspecting multiple Gmail accounts, including newly created ones, it was confirmed that three specific smart feature settings were automatically turned on.
The first setting, “Turn on smart features in Gmail, Chat, and Meet,” permits Google to utilize content from these applications to deliver personalized experiences and smart functionality. The second, “Smart features in Google Workspace,” applies to business and education accounts, enabling features like calendar events generated from Gmail emails and personalized search results. The third, “Smart features in other Google products,” extends this personalization to services such as Google Maps, Google Wallet, and the Gemini app.
A class action lawsuit filed in federal court in California alleges that Google secretly enabled Gemini to access private communications from Gmail, Chat, and Meet users without proper consent. The suit claims this may violate state privacy laws, arguing that Google made it challenging for users to opt out of the data tracking. While Google’s recent statements appear to counter the suit’s primary allegations, the automatic activation of these settings continues to raise privacy concerns.
For users who prefer to disable these features, the process is straightforward. On the desktop Gmail site, click the gear icon, navigate to “See all settings,” and under the General tab, locate the Smart features section. Uncheck the box for “Turn on smart features in Gmail, Chat, and Meet.” Then, select “Manage Workspace smart feature settings” and toggle off the options for Google Workspace and other products.
In the Gmail mobile app, access Settings through the menu, then go to Data privacy on iOS or your Google account name on Android. Disable the main Smart features switch, then enter the Google Workspace smart features menu to turn off the remaining two options. Be aware that disabling these settings will limit or remove functionality like Smart Compose and Smart Reply, though core Gmail operations will continue unaffected. Ultimately, the decision involves weighing personalized convenience against data privacy preferences.
(Source: ZDNET)
