Google expands Gemini AI use, privacy rules and smart‑home features

Google has updated its privacy policy to allow images, audio and other multimedia uploaded to its services to be used for training Gemini, the company’s generative‑AI model. Users can opt out through the Search History and Service Personalization settings, disabling options such as “Save multimedia content” or turning off the Search History entirely.

The firm also began rolling out Gemini for Home in early‑access mode, bringing a more conversational AI assistant to Google Home devices and smart displays. The upgrade promises natural‑language queries, smarter automations and faster handling of lights, thermostats, cameras and other connected devices. A new Google Home speaker featuring Gemini Live was reviewed, highlighting its compact size, improved bass and a built‑in switch to mute the microphones.

Gemini is being integrated into Android Auto, enabling drivers to use a conversational assistant for navigation, message summarisation, email lookup and other tasks while keeping eyes on the road. Google collected user feedback on the Gemini app, publishing the top ten requests: more reliable Google Workspace integration, better tool‑calling, improved chat organization, expanded custom skills, deeper research features, watermark removal, message editing, voice‑dictation accuracy, mobile‑app scrolling fixes, and unchanged celebrity‑likeness guardrails.

These changes span multiple Google products, aiming to make Gemini a core component of the company’s ecosystem while giving users control over data usage.