Is Gmail training Gemini AI based on your emails? Google responds to viral claim

In response to widespread speculation, Google has addressed claims that it is utilizing Gmail data to train its Gemini AI. A recent report from Malwarebytes suggested that changes in Google’s settings could lead to automatic training on users’ emails, prompting a wave of backlash on social media as users expressed their frustration over potential unauthorized use of their personal data. Google took to social media to set the record straight, emphasizing that no changes had been made to user settings and that Gmail content is not being used for Gemini AI training. The company stated, "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 clear if we make changes to our terms & policies." Following the uproar, Malwarebytes revised its original report, confirming that Google does not train its AI on users’ emails. The confusion stemmed from how Google had recently communicated its settings, which led many, including Malwarebytes, to mistakenly believe that Gmail content would be included in AI training. The tech firm noted, "The settings themselves aren’t new, but the way Google recently rewrote and surfaced them led a lot of people to believe Gmail content might be used to train Google’s AI models, and that users were being opted in automatically." This incident is not the first time Google has faced scrutiny regarding Gmail this year. In September, the company refuted claims that a leak had compromised 2.5 billion Gmail accounts, stating that such reports were incorrect and contributed to unnecessary panic. While Google has clarified that it is not using Gmail data for AI training, it does automatically train its AI models based on user interactions with its chatbot, Gemini. To prevent this, users must disable the Gemini Apps Activity setting in their Google account. This practice isn't unique to Google; competitors like Anthropic and Meta have indicated similar training protocols based on user interactions with their chatbots.