With a new “Help improve Assistant” option, a handful of users are now able to “save audio so speech technologies can learn over time.” This new capability also appears in the Google app’s General preferences.
Google Assistant is Improving its Accuracy
According to Google, “Your audio recordings stay on your device while a privacy-preserving technology combines information from you and many other participants to help Assistant learn over time and develop better smart features.” The purpose of the new option is to bring improvement in the accuracy of the “Hey Google” voice command in scenarios such as the ones described by Google:
On rare occasions, misactivate, or trigger when you didn’t say “Hey Google.” In other cases, miss, or not trigger, when you did say “Hey Google.”
In case you want to enable the new option, your voice recordings from near activations will not be saved on servers of Google as they will be processed locally. Moreover, all voice recordings are deleted after they have been processed by the algorithm. Google is continuously trying to improve the assistant in every possible way and this new addition is appreciable. It is important to mention here that Google is already using federated learning on Gboard and the Health Studies app. Recommended Reading: Google Assistant for Android Tests Memory Feature
