Google has always been banking on AI and machine learning to help improve its process in lieu of manpower. Machine learning, however, always requires more data to learn from, and there is perhaps no better source of millions of pieces of data than the millions of Android users out there. One such example is in optimizing how apps are downloaded and installed, a federated learning feature that Google has now started rolling out to Android users.
Google seems to be on a spree pushing a new style of machine learning that, at least according to the company, manages to preserve user privacy while still harvesting pieces of their data. There has, of course, been some controversy around its new Federated Learning of Cohorts or FLoC but that isn’t the only place it has been pushing a sort of crowdsourced form of learning. It is using the technique to improve Google Assistant’s hotword detection and now it’s doing something similar for Google Play Store’s app installation process.
Revealed just last month, the new App Install Optimization feature leverages what Google Play Store is already able to do on Android anyway. It basically learns which part of an app a user goes to first after installing it. It gathers that data from other users which it uses to improve how apps are packaged, downloaded, installed, and run to cover the most common use cases.
XDA has now received a tip that App Install Optimization has started to roll out to some Android users. A notice pops up when opening Google Play Store, informing them of this new feature. A new setting is also added so users can turn that on or off as desired.
It should be noted that App Install Optimization is enabled by default when it arrives. Google promises that no personal information is ever gathered and that the feature doesn’t even look at any other app. Whether that’s enough to satisfy privacy concerns, however, remains to be seen.