Google Denies Using Gmail Content To Train AI Models

JAKARTA Not long ago, many reports stated that Google used messages in Gmail to train their Artificial Intelligence (AI) model. This news was firmly denied.

Previously, Malwarebytes reported that Google had changed its AI training policy. The company is said to be using messages and all user attachments within the Gamil to train its AI model, Gemini.

Reports circulating also state that the only way to eliminate AI access to messages and Gmail attachments is to turn off 'smart features' or AI-based such as spelling checks. This report went viral on social media.

Google spokeswoman Jenny Thomson firmly denied the claim and stated that the widely circulated report was 'misleading'. Thomson added that the Smart Feature in Gmail had existed for years without a policy change.

"We haven't changed anyone's settings, Gmail's Smart Features have been around for years and we don't use your Gmail content to train our Gemini AI model," Thomson said.

The only problem with this feature is a system that can accidentally reactivate its features. Users are asked to recheck their settings because the Smart Feature can turn back on accidentally.

Although it does not change its main policy, Google briefly updated personalization settings for Smart Features last January. With this update, users can disable these features separately for Google Workspace and other Google products such as Maps and Wallet.

If activating this feature in Workspace, users agree that Google can use their content and activities to personalize the experience in Workspace. However, according to Google, this has nothing to do with the use of content to train AI.