Google has officially denied viral allegations that Gmail uses personal emails and attachments to train its Gemini AI model. A company spokesperson confirmed that user email content is not used for Gemini AI training.
Key Takeaways
- Google denies using Gmail emails to train Gemini AI
- Smart features like spellcheck and predictive text are separate from AI training
- Confusion arose from Malwarebytes report about email analysis
- Class action lawsuit filed alleging unauthorized tracking
Official Statement from Google
Google spokesperson Jenny Thomson told The Verge: “We do not use your Gmail content for training our Gemini AI model.” She clarified that Gmail’s smart features—including spellcheck, predictive text, package tracking, and calendar integration—have existed for years and don’t grant Google access to personal emails for AI training purposes.
Source of Confusion
The controversy began when Malwarebytes published findings suggesting Google was analyzing private emails and attachments to improve AI tools like Smart Compose and Smart Reply. Google introduced separate toggle options in January for Workspace smart features and other products like Maps and Wallets, which may have caused user confusion about what data is being used for what purpose.
Smart Features vs AI Training
Google explains that smart features analyze content locally or in limited ways to “personalize your experience across Workspace,” but emphasizes this is fundamentally different from feeding email correspondence to Gemini AI. However, the company’s explanation hasn’t fully resolved concerns.
Legal Action
On November 11, a class action lawsuit was filed against Google alleging the company tracks consumers’ communications without consent after allegedly secretly enabling Gemini AI for all users.
