The ability to understand voices and execute the respective command is probably the most crucial feature of any modern-day voice assistant, including Google Assistant. While these assistants are really good at understanding the voice commands, at times they struggle to understand things and resulting in vague replies.

It appears that Google is working on making things better. 9to5Google has spotted new capabilities for Google Assistant in the Google app APK teardown. According to the report, the APK teardown of the app includes strings about ‘personalised speech recognition’ for Assistant.

The app update is set to bring better voice recognition by storing the voice data on users’ devices and using it to optimise the speech recognition to adapt the tone of voice and the way users talk.

As per the report, the new voice recognition will work alongside the Assistant’s federated learning model introduced in March 2021. The APK teardown also reveals that users can turn off the feature and turn it off with the following message “If you turn this off, your Assistant will be less accurate at recognizing names and other words that you say frequently. All audio used to improve speech recognition for you will be deleted from this device.”

While the existing speech recognition system where the assistant asks users to repeat a couple of phrases at the beginning and then use it to recognise all the future commands, the added feature will offer definitely improve the vast difference in accent, speech and tone of voice.

Do note that the feature has been found under the APK teardown and it isn’t live as of now. However, it is expected that the company will roll out it out with a future update of the Google app.