Google is addressing recent viral social media claims and articles that suggest it has altered its policy regarding the use of Gmail messages and attachments for training artificial intelligence models. These claims state that users must disable “smart features,” such as spell checking, to opt out of this data usage.
In response to these claims, Google spokesperson Jenny Thomson informed The Verge that the reports are misleading. Thomson asserted that Google’s settings have not changed, stating that Gmail Smart Features have been available for years and clarifying that the content of users’ Gmail accounts is not utilized to train the Gemini AI model.
Although users might want to revisit their preferences, it has been noted by a Verge staff member that they had previously opted out of certain Smart Features but found themselves re-enrolled in them. Earlier this year, Google updated its smart feature settings to give users the ability to disable features across Google Workspace and other Google services—like Maps and Wallet—independently.
When Gmail’s smart features are enabled, they provide various functionalities, such as tracking orders and integrating flight information into calendars. According to Google’s settings page, enabling these features indicates that users consent to having their content and activity used to personalize their experience within Google Workspace. However, the company maintains that this consent does not imply that email content is employed for AI training purposes.
Source: https://www.theverge.com/news/826902/gmail-ai-training-data-opt-out