Lori Greiner warns of Gmail AI’s default email scanning; Google has urgently updated it

According to a report by Forbes on May 7, Lori Greiner, a star from Shark Tank, posted on Instagram over the weekend to warn that Gmail AI defaults to scanning emails. In response, Google has made an update, emphasizing that Gemini will not use users’ email to train AI models, and that it will not retain any data after processing.

Complete steps to turn off Gmail AI scanning in 30 seconds

To disable Gmail’s AI scanning feature, go to the “Smart features and personalization” section on the Gmail settings page, and turn off the following two settings:

Smart features: controls Gemini’s AI assistant tools, including smart replies, email summaries, and smart search

Personalization: controls whether Gmail allows other Google products to access your email data

Important note: After disabling, you will not be able to use any new AI upgrade features. Google states that if you want to re-enable after disabling, you need to perform specific reconfiguration steps and cannot switch back on at any time like with ordinary features.

Google’s official guarantee: Gemini will not retain your emails

Google provided two core guarantees with this update: first, this AI upgrade “will never use user email content to train its AI models”; second, Gemini in Gmail is “designed to run securely inside the inbox, handling only the content the user requests, and then leaving the inbox afterward, without retaining any data.”

Google also emphasized that these features are optional for all users, but they are turned on by default—if users want to disable them, they need to proactively go to settings to change them, and the system will not prompt users automatically.

More worth attention: the “Personal Intelligence” feature is more intrusive than email scanning

Although Lori Greiner’s viral post focuses on Gmail email scanning, the concurrently released “Personal Intelligence” update is more worthy of careful evaluation. This feature scans users’ various data across multiple Google platforms, builds a comprehensive situational model of users’ daily lives, and allows users to choose whether to connect other applications to Gemini. Its data collection scope goes far beyond simple email scanning, and it has already been made available to millions of users.

In addition, around the same time, privacy researchers revealed that the Chrome browser silently downloads about 4GB of the Gemini Nano model without informing users. Taken together, these incidents point to systemic gaps in Google’s proactive notification mechanisms when deploying AI features at scale.

Frequently asked questions

Will Gmail’s Gemini AI feature use my emails to train Google’s AI models?

Google clearly guarantees that this Gmail AI update will not use user email content to train AI models. The design principle for Gemini in Gmail is: it only processes relevant emails when the user makes a specific request, retains no data after processing is complete, and Gemini’s computation will not be sent to the cloud.

After turning off “Smart features” and “Personalization,” what functions will I lose?

After disabling, you will lose Gemini-powered smart reply suggestions, automatic email summaries, smart search, and the personalization service that allows other Google products to access your email data. According to Google, if you want to re-enable after disabling, you need to perform specific reconfiguration steps and cannot switch back directly.

What is the connection between the incident of silent downloading 4GB Gemini Nano in Chrome and this Gmail update?

The two are separate incidents from the same period by Google that are thematically related. Chrome’s Gemini Nano is mainly used for on-device security functions (such as scam detection), while Gmail’s Gemini features focus on AI assistance in the inbox. Together, they have sparked widespread skepticism about Google deploying AI features broadly on devices and accounts without sufficiently notifying users.