A recent wave of confusion arose online when it was claimed that Google used Gmail messages to train its Gemini AI. This misinformation sparked widespread concern across social media, prompting many users to question Google’s data privacy practices. The confusion began after MalwareBytes published an article suggesting that changes to Gmail settings could permit user messages to contribute to Gemini’s training.
In response, Google quickly denied these claims. Spokesperson Jenny Thomson clarified that the reports stemmed from a misunderstanding regarding Gmail’s Smart Features, which have been in place for years and require user consent. She emphasized that Gmail messages are not used to train Gemini, which is developed on a completely separate dataset.
MalwareBytes later updated their article to address the confusion, acknowledging that while Gmail scans messages for features, this does not imply content is utilized for training generative models like Gemini. Google’s Smart Features are optional, ensuring user control over data settings.
Source link
