Google integrates its Gemini Nano AI model into Chrome on the desktop

At the Google I/O 2024 developer conference on Tuesday, Google announced that it is integrating Gemini Nano, the smallest of its AI models, directly into the Chrome desktop client, starting with Chrome 126.

According to the company, this will allow developers to use the on-device model to power their own AI features. Google itself plans to use this new feature to power features like Workspace Lab’s existing “help me write” tool in Gmail, for example.

The company says it’s recent work on WebGPU and WASM support in Chrome that allows these models to run at reasonable speed on a wide range of hardware.

In a briefing ahead of Tuesday’s announcement, Jon Dahlke, Google’s director of product management for Chrome, noted that Google was in talks with other browser vendors to also enable this feature – or a similar feature – in their browsers.

“We have started collaborating with other browsers and will open an initial preview program for developers,” Dahlke wrote in Tuesday’s announcement. “With webGPU, WASM, and Gemini built into Chrome, we believe the web IS ready for AI. “

I doubt, however, that most of Chrome’s competitors want to bet solely on Google’s AI models. What makes more sense is to allow browsers – and developers – to run whatever model they want. Google would clearly choose to run Gemini for its apps, but these models are small enough to allow developers to choose which one they want for their apps.

Google’s bet, however, is to enable a number of high-level APIs in Chrome to translate, subtitle and transcribe text in the browser using its Gemini models.

“To deliver this feature, we refined our most efficient version of Gemini and optimized Chrome,” Dahlke said during today’s developer keynote at I/O. “We now want to give you access to Gemini templates in Chrome. Our vision is to bring you the most powerful AI models in Chrome to reach billions of users without having to worry about rapid engineering, fine-tuning, scale and cost. All you have to do is call a few high-level APIs: translate, subtitle, transcribe. This is a big change for the web and we want to make it happen.

For developers, Google is now also using the built-in Gemini Nano model to power some new features in the Chrome DevTools console. With this, Chrome’s developer tools can now explain errors and provide debugging solutions directly in the console.

We are launching a newsletter on AI! Sign up here to start receiving it in your inboxes on June 5.