Gemini Live Gets Model Update for Enhanced Conversations
Google is rolling out a model update to Gemini Live, promising more dynamic and engaging conversations. This update leverages the latest Gemini model, improving comprehension of diverse languages, dialects, and accents within a single chat, and enhancing translation capabilities. DeepSeek's advancements in AI are likely contributing to these improvements.
The update also paves the way for screen sharing and live video streaming within the Gemini app, features previously demonstrated with Astra. These enhancements align with the capabilities of the Gemini 2.0 Multimodal Live API, which processes text, audio, and video input while outputting text and audio. For more on Gemini 2.0, see this article about iOS 18.4's new features and improvements.
Data Privacy and Storage Changes
With this update, Gemini will store audio, video, and screenshares in Gemini Apps Activity. This change reflects an expansion from the current practice of storing only Live chat transcripts. Users can manage and delete their activity data according to their auto-delete settings. Concerns about data privacy are addressed in the FBI's warning on AI-powered phishing.