Particle.news
Download on the App Store
14 ARTICLES
·
2w ago

Google Rebuts Claims Gmail Trains Gemini as Malwarebytes Retracts Post

Google says viral reports misread longstanding Smart Features that personalize Gmail without contributing emails to Gemini’s training data.

Google has been quietly rifling through users' personal Gmail accounts to train its AI models on their emails and even attachments.
Photo: Stephen Phillips
blank
Image

Overview

  • Google’s official Gmail account said it has not changed user settings and does not use Gmail messages or attachments to train its Gemini AI.
  • Malwarebytes updated its initial report, acknowledging that its interpretation of Gmail settings was mistaken and that the features in question are not AI training.
  • Gmail’s Smart Features scan email content for personalization such as spam filtering, categorization, and writing suggestions, which Google and multiple outlets distinguish from model training.
  • Some users report Smart Features appearing enabled or re-enabled, and guidance to manually turn off Gmail and Workspace smart features has circulated, with reports noting stricter default protections in the EEA, UK, Switzerland, and Japan.
  • A proposed class-action filed Nov. 11 in federal court in San Jose alleges Google secretly enabled Gemini access to Gmail, Chat, and Meet in October, a claim Google disputes as the case proceeds.