Google has launched a pilot project to integrate Built-in AI, preparing to introduce Gemini Nano directly into Chrome, opening up browser APIs in the LLM group for additional features. Websites will now be able to utilize LLM on the web, such as summarizing content or answering questions within the website itself, even to the extent of acting as a user’s content writer assistant or spell checker.
Many websites are already running artificial intelligence models on the web, but the daily use of such models requires users to download large files, often in hundreds of megabytes in size for each website, and may have limited access to machine features like GPUs or NPUs. Direct browser APIs will enhance optimization.
The current use of AI on Chrome is still experimental to find suitable APIs. It is anticipated to introduce Prompt API for normal usage and Fine-tuning API for specific model utilization.
Source: Chrome for Developers
TLDR: Google is testing Built-in AI with Gemini Nano on Chrome, providing direct browser APIs for advanced features like content summarization and question answering within websites, with plans to introduce more APIs for regular and specific model operations.
Leave a Comment