Google Denies Using Gmail to Train Gemini as Users Question Privacy Settings
Social media has been buzzing with speculation that Google is secretly feeding Gmail content into its Gemini AI model. Over the past week, posts have gone viral alleging that emails, attachments, and personal conversations are being tapped for AI training—fuelled further by claims that Google quietly altered privacy settings without user consent.
But the company says the rumours are simply incorrect.
In a statement to a famous publication, Google spokesperson Jenny Thomson dismissed the claims, stressing that the fears circulating online are misplaced. “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,” she said.
Despite this clarification, the online uproar has prompted many Gmail users to take a second look at their settings, unsure of what has actually changed—and what hasn’t.
Why Users Are Confused About Smart Features
The controversy appears to trace back to an update Google introduced earlier this year. Since January, users have been able to manage “smart features” independently across Google services. This means options in Gmail or Workspace—like auto-draft suggestions, email categorisation, and smart reminders—can be disabled without affecting features in Maps, Wallet, or other applications.
However, as the popular publication noted, some users noticed that their previously disabled settings had been quietly turned back on. The extent of this glitch remains unclear, but it was enough to spark suspicions that Google might be re-enabling these features to expand its AI data sources.
These tools do more than simple spell checks—they automate flight updates in your calendar, highlight package tracking, and summarise emails. When turning them on, Gmail informs users that the service will rely on “Workspace content and activity to personalise your experience.”
While that phrasing may sound unsettling to some, Google maintains that such personalisation stays strictly within the user’s account environment. It helps Gmail function more efficiently for individuals, but does not feed data into Gemini’s training systems, the company argues.
Spotlight on Gemini 3 Raises Questions
The renewed scrutiny comes at a time when Google is heavily promoting Gemini 3, its most advanced AI model yet. The system is designed to process and interpret information with what the company describes as human-like depth.
CEO Sundar Pichai highlighted Gemini 3’s enhanced long-context reasoning and its ability to analyse everything from academic papers to video lectures. He wrote that the new model “is state-of-the-art in reasoning, built to grasp depth and nuance, whether it's perceiving the subtle clues in a creative idea, or peeling apart the overlapping layers of a difficult problem.”
Pichai added that Gemini 3 “is also much better at figuring out the context and intent behind your request, so you get what you need with less prompting.”
As Google pushes the boundaries of what AI can understand, user anxieties about how their personal data is used will inevitably grow. For now, Google insists your inbox is not fueling its AI evolution—but the debate shows how closely users are watching every update.