How to Stop Google From Using Your Gmail for AI Training – Step-by-Step Privacy Fix

How to Stop Google From Using Your Gmail for AI Training – Step-by-Step Privacy Fix

If you use Gmail, there’s a good chance your emails are being used to train Google’s AI models. A recent report from TechRepublic highlights that the company’s default privacy settings may allow this without clear consent, and a lawsuit filed in early 2026 claims that the option is effectively hidden from users. With 1.8 billion active Gmail accounts, that’s a lot of personal correspondence feeding into machine-learning systems.

Here’s what’s really going on, why you might want to check your settings, and—most importantly—how to turn it off if you prefer your private messages stay private.

What happened

In January 2026, Google was sued over what the plaintiffs described as a “hidden AI training setting” in Gmail. According to court documents and reporting by TechRepublic, the company had been using user emails to improve its generative AI products, including features like Smart Reply, summarization, and experimental chatbots. The lawsuit argues that Google did not adequately inform users or obtain proper consent before enabling this data use.

Separate coverage also points to a broader concern: research published in March 2026 showed that AI “agents” can leak personal data if they are trained on sensitive information like email content. The combination of these events has pushed email privacy back into the spotlight.

Why it matters

Your Gmail inbox often contains intimate or sensitive material: medical appointments, financial statements, personal conversations, and work communications. If those messages are used to train AI models, they could potentially be reconstructed or inferred by other users of the same AI system. Even if Google applies anonymization, the risk—however small—is non-zero, and many people prefer their data simply not be used for this purpose at all.

The important nuance is that, as of mid-2026, Google says this setting is optional. The catch is that it may be turned on by default for some accounts, especially those that opted into certain AI features without reading the fine print.

What readers can do

Here is a step-by-step guide to check and disable the AI training setting for your Gmail account. These steps work on both web (desktop) and the Gmail mobile app.

1. Open your Google Account settings.
On the web, click your profile picture in the top-right corner of any Google service and select “Manage your Google Account.” On the Gmail mobile app, tap your profile icon, then “Manage your Google Account.”

2. Go to “Data & Privacy.”
In the left menu (or the tab bar on mobile), find and tap “Data & Privacy.”

3. Scroll to “AI and personalization.”
This section used to be called “My Ad Center,” but in 2025 Google moved and renamed several privacy controls. Look for a heading that says “AI and personalization” or “AI training.” If you don’t see it, use the search bar at the top of the page: type “AI training.”

4. Find the toggle for Gmail.
Under “AI and personalization,” you’ll see a list of services including Gmail. The description may read something like “Allow Google to use your Gmail content to improve AI models.” If the toggle is blue (on), it means your emails are being used.

5. Turn it off.
Tap the toggle to switch it to gray (off). Google will ask you to confirm your choice. It may also warn that some features, like “Smart Reply” or “Summarize this email,” might stop working. That’s true—if you disable training, Google can’t use your specific data to personalize those AI features. They may still work in a generic way, but results could be less tailored.

6. Double-check other services.
While you’re in the “Data & Privacy” section, also review settings for YouTube, Google Photos, and Google Workspace (if you use a business account). The same AI training toggle might be enabled for those services separately.

If you want to be thorough, you can also limit ad personalization and activity tracking under “My Activity” and “Ad settings.” Those controls are separate and don’t affect AI training, but they reduce what Google collects about your overall behavior.

Keep in mind

Google does not currently allow you to delete the training data it has already collected from your emails, only to stop future use. If privacy is a priority, you might consider using an email provider that does not mine your messages for AI purposes. Proton Mail and Tutanota, for example, offer end-to-end encryption and have publicly stated they do not train AI on user content.

That said, for most people, simply flipping the toggle is a quick fix that takes less than a minute. With 1.8 billion users, even a small percentage who now opt out adds up to a substantial reduction in personal data flowing into AI models.

Sources

• TechRepublic, “1.8 Billion Gmail Users May Want to Check This AI Privacy Setting,” May 2026
• TechRepublic, “Google Faces Gmail Lawsuit Over Hidden AI Training Settings,” January 2026
• TechRepublic, “‘Agents of Chaos’: New Study Shows AI Agents Can Leak Data, Be Easily Manipulated,” March 2026