Google Reiterates That It Will Not Train AI With Your Gmail Data

One of the enduring myths in personal technology is that Google is reading your Gmail messages. But this is a bigger issue than ever in this AI era. And so Google is explaining how it handles Gmail, Gemini, and privacy.

“With so many new AI tools and features, it’s understandable if you have some questions about your privacy in Gmail,” a new post to Google’s The Keyword blog reads. “[But] your inbox is your business.”

According to Google, it does not train its foundational AI models, including Gemini, on your personal emails. When you grant Gemini access to Gmail, it will only perform isolated tasks, like summarizing a long email message, and nothing more. “What’s in your inbox stays private, even if you use Gemini to help you,” Google says.

Furthermore, Gemini in Gmail doesn’t retain any of the data it uses to help you with your email. Google says that Gemini works securely in your Gmail inbox, processes what it needs to perform the tasks you require, and then it leaves your inbox and doesn’t retain any data it used.

“We take your security and privacy very seriously,” Google vice president of product for Gmail Blake Barnes says. “Gemini and Gmail work together to keep you safe.”