Google is rolling out AI enhancements across its Android devices, impacting over 2 billion Gmail users and raising privacy questions. The company’s AI, named Gemini, will access content across apps such as Phone, Messages, and WhatsApp, even if users disable Gemini Apps Activity.
Initial communications from Google caused confusion, suggesting AI would review sensitive messages regardless of settings. Google has since clarified that while Gemini can assist with tasks like sending messages and making calls with Apps Activity off, it does not review or use chat data to train AI models under this setting.
Previously, using Gemini with messaging apps required keeping Apps Activity enabled, which allowed data storage and training—a significant privacy concern now addressed by the update. However, interactions are still saved temporarily for 72 hours even when Apps Activity is disabled.
Despite this clarification, privacy concerns persist. Reports highlight that deep integration of AI into personal communications may expose sensitive data such as call logs and private chats, triggering user distrust.
Similar AI features are emerging in other platforms like WhatsApp, which now offers AI-generated message summaries. This broad adoption underscores the ongoing balance between convenience and privacy, especially for less aware users, including younger demographics.
Google is also introducing Gemini in educational settings, sparking debate about AI’s role in schools and its long-term effects on learning. According to TechRadar, student data used with Gemini for Education is protected under strict policies, ensuring it is not employed for AI training or human review.
A recent report from Incogni ranks Google’s AI privacy practices more favorably than some competitors, but warns users remain generally uninformed about the complex privacy implications of widespread AI integration in devices and apps.
As AI becomes embedded in daily tools, experts urge heightened awareness of data handling practices and ongoing privacy risks associated with large language models and generative AI.