Google Integrates Gemini Nano AI Model into Desktop Chrome

Home AI Projects Google Integrates Gemini Nano AI Model into Desktop Chrome
Google Integrates Gemini Nano AI Model into Desktop Chrome

At the Google I/O 2024 developer event on Tuesday, Google revealed its plans to integrate the smallest of its AI models, the Gemini Nano, into the Chrome desktop browser starting with Chrome version 126.

 

According to the company, this integration will allow developers to use the built-in model to control AI functions. Google intends to use this innovation to improve features like Workspace Lab’s Help Me Write feature in Gmail. The company credits recent developments in Chrome’s WebGPU and WASM support for allowing these models to run efficiently on a variety of hardware.

 

During a briefing before the announcement, John Dahlke, Google’s director of product management for Chrome, mentioned ongoing discussions with other browser vendors to bring this or a similar feature to their browsers. Dahlke said in a statement: “With the implementation of WebGPU, WASM, and Gemini in Chrome, we are ready to declare the Internet ready for AI.”

 

While it’s unlikely that Chrome’s competitors will rely solely on Google’s AI models, a more practical approach would be for browsers and developers to freely choose and run their preferred models. Google may prefer to use Gemini for its services, but the small size of the models should give developers flexibility in choosing the right ones for their applications.

 

Google is also introducing high-level APIs in Chrome to perform functions such as text translation, subtitling, and transcription with its Gemini models, Dahlke emphasized during the keynote. He explained: “We’ve honed our most optimized version of Gemini and tweaked Chrome for optimal performance. We’re committed to making Gemini models available in Chrome to bring world-class AI capabilities to millions without the complexities of rapid design, fine-tuning, and cost and power management. Just use a few basic APIs – translation, subtitling, transcribing. This is a significant transformation for the Internet, and our goal is to improve it.” In addition, the new built-in Gemini Nano now provides additional features to the Chrome DevTools Console, allowing it to automatically decipher errors and offer debugging tips right in the console.

allix