Google Says It Doesn’t Use Your Emails to Train Its Artificial Intelligence.
Last week, tech headlines were filled with an alarming claim: according to some social media users and even some news outlets, Google secretly allowed Gmail users to share their emails with Gemini. The goal? To train Google’s artificial intelligence. While the controversy generated quite a bit of buzz, it seems the main discussion began with this provocative X post , typed in all caps:
This tweet is currently unavailable. It may be loading or has been removed.
According to the author, the only solution is to disable a setting you’ve likely never heard of: “smart features.” The statement claims that “smart features” were enabled in two places in your account: one for Gmail, Chat, and Meet, and the other for enabling them in the broader Google Workspace suite of apps (assuming you have a Workspace account). The “irrefutable proof” is a small section of Google’s privacy policy that states the company does not use your Workspace data to train its AI models without permission . Taken at face value, the claim is that Google enabled these settings, thereby gaining your “permission” to begin using your data to train its AI models.
By this morning, the post had garnered nearly 150,000 likes on X, but the accusations had spread across the internet. Even Malwarebytes published an article explaining the situation and urging users to disable these settings to prevent their posts from being included in Google’s AI training system. However, as it turns out, Google isn’t actually doing this.
What are “smart features”?
On Friday, I contacted Google to see if the company had any comment on these statements. In response, I received the following from a Google spokesperson:
These reports are misleading—we haven’t changed anyone’s settings, Gmail’s intelligent features have been around for years, and we don’t use your Gmail content to train our Gemini AI model. Finally, we always act transparently and forthrightly when making changes to our terms of service and policies.
Indeed, “smart features” have been around for years. Lifehacker first covered the topic back in 2020 , nearly three years before Google launched Bard (now Gemini). To the plaintiffs’ credit, smart features do use your data to improve Google features, but not to mass-train AI models. Instead, smart features use your data to customize your specific Google experience. In Gmail, this might mean learning how you write to generate more relevant smart replies; in Search, it might mean suggesting keywords based on previous searches; in Maps, it might mean showing restaurant reservations; in Wallet, it might mean recommending show tickets based on your interests. Smart features also power core Google features like spam filters and categorization.
This doesn’t mean you have to use smart features: if you’re personally uncomfortable with Google collecting data to personalize your experience, you can turn them off. But it’s important to understand what you’re turning on or off: even if someone claims these features do something, that doesn’t mean they’re true. Disabling smart features is fine, but be aware that doing so won’t affect how Google handles your personal messages or how it trains its AI models for all users. It may also impact how apps like Gmail work: you might not be able to sort emails by category or block certain spam from reaching your inbox.
How to disable smart features
If you want to disable Smart Features, it’s easy to do. The settings are located in the Gmail web app: click the settings gear icon, then select “View all settings.” Scroll down and turn off the toggle next to “Smart Features.” For the changes to take effect, you’ll need to agree to the “Turn off and restart” option in Gmail.
If you have a Workspace account, you’ll have access to another settings page. Click “Manage Workspace intelligence settings,” then uncheck the boxes next to “Intelligence in Google Workspace” and “Intelligence in other Google products.” Then click “Save.”
Please note: these settings control a number of smart features in your Google products. Therefore, you may notice some features are missing in the future. To undo your changes, return to these settings pages at any time and turn the toggles on.
Google is not safe
Since the furor peaked last week, the dust has settled. Malwarebytes published a correction to its original article , reprinting the same statement from a spokesperson I received. The site claims that Google’s vague rewrite of its intelligence settings contributed to the confusion, but this setting doesn’t appear to collect users’ private messages for Gemini’s training.
However, this doesn’t mean Google is completely innocent here. While this specific claim may be incorrect, the lawsuit alleges that the company improperly transferred personal data of Gmail, Chat, and Meet users to Gemini without their permission. Google, in general, isn’t known for respecting user privacy, so it’s unsurprising that these allegations have received such widespread attention.
Malwarebytes also notes that while Google claims all these settings are optional, many users don’t remember enabling them, further exacerbating the confusion surrounding Smart Features. Frankly, I don’t recall enabling Smart Features in either my personal or work Gmail accounts, although it’s entirely possible I did. I only know that these settings are currently enabled, and while I’m considering disabling them, I’m not sure I’m ready to give up Smart Features entirely.