It’s a cool idea but voice recognition isn’t that high on a list of must have features for most mobile phone users out in the big wide mobile world, but comes in handy at times like when driving. So it’s nice to hear that AT&T is looking to bring true voice activated text entry to their mobile handsets via a web based speech recognition service.

AppleInsider reports that the research project leverages AT&T’s WATSON speech recognition engine, referred to as Speech Mashups to bring voice activation features to any mobile handset holding a high speed wireless data connection.

AT&T’s Speech Mashups service will work with AT&T’s 3G network to transmit voice to a remote speech recognition server, translate voice into computer language and the send commands back to the mobile handset or website.

However, Speech Mashups service is not designed to offer voice dialling or direct voice control to mobile handsets, therefore native apps need to be developed with specific code in order to take advantage of speech recognition.