Google Debunks Viral Gmail AI Training Rumors as False Alarm

Google is addressing claims circulating on social media asserting that the company altered its privacy policy to utilize Gmail messages for artificial intelligence training. In a statement to The Verge, Google clarified that these reports are “misleading” and confirmed that it does not use Gmail content to train its Gemini AI model. However, some users report being automatically re-enrolled in smart features they had previously opted out of, raising concerns about privacy.

The allegations gained traction through various platforms, including a viral Twitter thread and a report from Malwarebytes, suggesting that users need to disable certain “smart features,” such as spell-checking, to prevent their emails from being used for AI model training. Google spokesperson Jenny Thomson reiterated that there have been no changes to users’ settings and that Gmail’s Smart Features have been available for several years.

Despite Google’s denial, some users have noted unexpected changes in their feature settings, suggesting a possible alteration in how these settings are managed. This situation may stem from a January update that modified Google’s smart feature personalization settings, allowing for more granular control over Google Workspace features.

Gmail’s smart features include capabilities like automatic flight confirmations and package tracking, which require users to consent to their data being used for personalized experiences across Google Workspace.