Viral Claims Spark Privacy Concerns

Social media has recently been flooded with posts alleging that Google is mining emails and attachments to power Gemini. Many users speculated that Google quietly updated its privacy policy to allow the use of Gmail data for AI training, advising others to disable Gmail’s “smart features” to avoid being included.

Google strongly disputes these claims. Speaking to The Verge, company spokesperson Jenny Thomson 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.”

Despite the reassurance, the claims have prompted many users to recheck their Gmail settings.

What’s Actually Going On With Gmail’s Smart Features

Confusion stems from a January update that let users enable or disable smart features separately across Google services. For example, you can turn off smart features in Gmail, Calendar, and Docs while leaving them on for Maps or Wallet.

However, The Verge also reported that some users experienced a silent reset of these settings, re-enabling features they had previously turned off. While it remains unclear how common this issue was, it fueled speculation that Google was subtly pushing users toward data-sharing.

These smart features are more than typo fixes—they power email summarisation, order tracking, and automatic calendar updates. When enabled, Google Workspace explains that it uses your content to personalise your experience across Workspace.

Google stresses that this personalisation is local to the user, not used to train Gemini or any large-scale AI model. The company maintains that Gmail data improves Gmail—for you, not for Google’s AI.

Gemini 3 Takes Center Stage

The controversy comes as Google launches Gemini 3, its most advanced AI model yet. CEO Sundar Pichai describes it as capable of interpreting information with “human-like depth and nuance,” combining multimodal understanding, long-context reasoning, and agentic behaviors.

“Gemini 3 is state-of-the-art in reasoning, built to grasp depth and nuance—whether it’s perceiving subtle creative cues or unraveling overlapping layers of a complex problem,” Pichai wrote in a recent blog post.