Google Denies Training Gemini AI on User Emails Without Consent
Google has moved to dispel concerns circulating online alleging that its new Gemini AI model is being trained on user email data without permission.In an official statement provided to The Verge, Google spokesperson Jenny Thomson labeled the reports as “misleading,” asserting that the company has “not changed anyone’s settings,” and that Gmail’s Smart Features, which have been available for years, do not utilize user email content for Gemini AI training.
The claims, which gained traction on social media platform X, sparked outrage among users worried about the privacy of their personal communications. One user characterized the situation as “the largest consent manufacturing operation in history.” The allegations were even amplified by a post on the blog of antivirus software firm Malwarebytes, further broadening their reach.
A widely shared social media post falsely claimed that all Gmail users were automatically enrolled in a program allowing Google access to their private messages and attachments for AI model training. The post directed users to disable Gmail’s Smart Features – AI-powered tools designed to enhance productivity – to prevent this alleged data access. However, Google’s Workspace Privacy Policy clarifies that while data directly shared with Gemini, such as prompts entered into the application, may be retained and used for AI training, data from apps like Gmail, Google Docs, and Sheets is not automatically accessed or used for AI training. The policy does state that Gemini can access Workspace data if a user explicitly directs it to do so, for example, by requesting the AI to proofread a document in Google Docs.
This is not the first instance of Google facing online misinformation regarding its services. In September,false reports went viral claiming Google issued a security warning urging all Gmail users to promptly change their passwords.
While the current claims regarding gemini and user email data appear to be unfounded,anxieties surrounding tech companies leveraging user data for AI training are understandable. in the European Union, where data protection laws are more stringent than in the United States, companies like meta and LinkedIn are increasingly announcing plans to utilize user data for AI advancement. Google itself has faced legal repercussions for data handling practices,agreeing to a $1.375 billion settlement in May 2025 after allegations of harvesting biometric data from Texans without their consent.
The incident underscores the growing need for openness and clear communication from tech companies regarding their AI practices, as well as the importance of critical evaluation of information circulating online. With Gmail boasting a user base of 2.5 billion people worldwide,even unsubstantiated claims can quickly escalate into widespread concern.
Why did this happen? The controversy stemmed from a false social media post claiming Google automatically enrolled all Gmail users in a program allowing access to their private messages for Gemini AI training. This claim was amplified by shares and a post on the Malwarebytes blog.
Who was involved? Google, its spokesperson Jenny Thomson, Gmail users, social media users on X, and the antivirus software firm Malwarebytes all played a role.
What was the claim? The central claim was that Google was secretly using Gmail user data to train its Gemini AI model without explicit consent.
How did it end? Google officially denied the claims, stating the reports were misleading and that user settings hadn’t changed.Google’s Workspace Privacy Policy was cited to clarify that Gmail data
