Google Expands Gemini AI to All Android Devices
Google is rolling out its Gemini AI to all Android devices with at least 6GB of RAM, significantly expanding its reach. The company also launched a Gemini Music Generator, which uses the Lyria 3 model to create 30-second songs from text or image prompts. The AI-powered features are now available in Google News for Android.
- The on-device model, Gemini Nano, comes in two sizes: Nano-1 with 1.8 billion parameters for low-memory devices and Nano-2 with 3.25 billion parameters for more powerful phones. This model is designed to run completely offline, without a network connection. - Google's hardware strategy centers on its Tensor Processing Unit (TPU), a custom chip optimized for AI and ML calculations that power features like computational photography. In contrast, Apple's A-series chips with the Neural Engine have shown superior performance in some machine learning benchmark tests. - The expansion lowers the minimum operating system requirement from Android 12 to Android 10 for the full Gemini assistant experience. The basic Gemini app is even more accessible, requiring only Android 9 and 2GB of RAM, although it is not supported on Android Go devices. - The Lyria 3 music generation model automatically creates lyrics based on the user's prompt and provides control over style and tempo. To address copyright concerns, Google states the model is designed for original expression, not mimicking artists, and watermarks all audio with its SynthID technology. - Beyond consumer devices, Google is positioning generative AI for industrial use, offering tools for a "supply chain advisor" to provide network visibility and recommend suppliers. - For developers, Google provides the ML Kit GenAI APIs, which use LoRA (Low-Rank Adaptation) adapters on top of the base Gemini Nano model to ensure consistent quality for features like summarization and proofreading across different devices and model versions. - The AI integration extends into the smart home, where Gemini allows for more natural, free-flowing conversations with Google speakers and displays without needing specific command phrases. - Within manufacturing, Google is deploying AI agents through its cloud platform to automate tasks like quality control, production planning, and predictive maintenance. 56% of manufacturing executives report their organizations are already actively using AI agents.
