For the past couple of days, news has been circulating online that Google is using your emails stored in Gmail to train its AI models. This claim has gained traction online thanks to viral social media posts and reports from even Malwarebytes that Google has introduced new AI features for Gmail to enable this behavior. Now, Google has decided to set the record straight on this topic.
In a post on X (formerly Twitter), the official Gmail account has refuted any claims around this topic, stating in no uncertain terms that it does not use your Gmail content to train Gemini, and that it has not modified anyone’s settings to enable this behavior either. The full post can be seen below:
Let’s set the record straight on recent misleading reports. Here are the facts:
• We have not changed anyone’s settings.
• Gmail Smart Features have existed for many years.
• We do not use your Gmail content to train our Gemini AI model.
We are always transparent and…
— Gmail (@gmail) November 21, 2025
Prior to this post, there were multiple allegations online that since you are opted in to Google’s “Smart Features” for Gmail automatically, the company can leverage that data to train Gemini. While the former part is indeed true, the latter is definitely not, according to Google. In fact, the company has a dedicated support article for customers about how they can disable Smart Features across all of its products.
This is not the first time in recent memory that Google has been the target of wild online claims. Just a couple of weeks ago, a major trove of data procured from breaches seemingly contained 394 million Gmail addresses. However, it was later confirmed that this data was not extracted through a Gmail hack and most of it had nothing to do at all with the service itself.
