New Chrome browser beta adds voice recognition API

A new beta version of Google's Chrome browser for PCs can run web applications controlled via user voice commands, so that people will be able to play computer games and compose documents through speech.

The beta of Chrome 25 features the new Web Speech API (application programming interface), which lets developers include speech recognition features in their applications, Google announced on yesterday.

"So, in the near future you'll be able to talk apps into doing all sorts of things," said Glen Shires, a Google software engineer and speech specialist.

Chrome 25 beta on Windows also will automatically disable some extensions that may have been "silently" installed on the browser by third-party applications, something Google first announced in December.

Google allowed the installation of these extensions via a Windows registry mechanism so that users could add third-party Chrome extensions they considered useful.

However, according to Google, the feature has been abused by many developers who have used it to install extensions without users' consent, resulting in potential slowdowns of the browser's operation.

Now, when a third-party application tries to install a Chrome extension, users will be prompted to either allow it or remove it from their computer.

Some previously installed extensions will be automatically disabled, and Chrome will display a one-time prompt giving users a chance to reactivate the ones they're interested in, according to Google.

Chrome 25 beta also includes some security enhancements for developers. For example, developers can now define a list of trusted content sources using the unprefixed Content Security Policy (CSP) HTTP header.

"The browser will only execute or render resources from those sources," said Eric Bidelman, a Google Chrome developer advocate.

This enhancement using CSP should reduce users' risk of suffering cross-site scripting and content injection attacks, according to Bidelman.