Gmail users are discovering they've been automatically opted into a controversial setting that allows Google to access their private emails and attachments to train AI models. Cybersecurity experts warn this hidden feature shares personal and work communications without explicit consent, raising significant privacy concerns among users who may not realize their data is being harvested.

Google's Smart Features, which include AI-powered tools like smart compose and automatic email categorization, are tied to this data collection practice. The company claims these features have existed for years and that Gmail content isn't used for training its Gemini AI model, but a proposed class-action lawsuit alleges otherwise. Users are accusing Google of secretly enabling AI access to their entire email history without proper disclosure.

Turning off this setting requires navigating through multiple menus in both desktop and mobile versions of Gmail. Users must manually opt out in two separate locations: first in the General tab under Smart Features, then in the Google Workspace settings. However, disabling these features also removes convenient tools like smart compose, spell-check, and automatic email filtering that many users have come to rely on.

This controversy highlights the growing tension between AI development and user privacy. As companies race to train their models, users are increasingly finding themselves opted into data-sharing arrangements they never agreed to. The incident underscores the importance of regularly reviewing privacy settings across all digital services.

Quick Fact: A proposed class-action lawsuit alleges Google secretly enabled AI access to users' entire Gmail history.