Federated Learning: Google Wants Your Data

Google's going to present themselves as a responsible force for good at Google IO 2018. They'll be rolling with a message that centers on not just the responsible handling of data, but the measured, reasonable use of smart devices. They're going to do this in several ways – one of which is with the concept "Federated Learning."

According to Google, there's a good way to make your smartphone smarter without giving up your privacy. In the very recent past, Facebook's suffered in a very public way due to their shocking lack of oversight when it comes to personal data and sharing without reasonable consent. Google's Federated Learning concept is such that your smartphone can benefit from the data of countless other smartphones without the submission of personal data.

ABOVE: "Making every phone smarter with Federated Learning" as published by Google on May 7th, 2018, less than 24 hours before the official start of Google IO 2018. We're expecting a collection of similar re-brandings and new initiatives throughout the week.

The concept here is not new. Google's been working with this kind of data flow for years. Without it, they'd have been unable to create the top-notch voice recognition models they have now for Google Assistant. They'd have Search Results that aren't nearly as well curated as they are today.

In short: without Federated Data (in one form or another), Google would have had to have done all this work on their own, one case at a time. And the results wouldn't be nearly as good as they are today. If you'd like to see all the data Google has on you – straight from you, not necessarily all used for Federated Data-like programs – have a peek at Google Takeout

The big deal this week is making the most of Google's current position on privacy and responsible usage of data. Google has the opportunity to pull ahead of the competition in Facebook, Apple, Microsoft, and Amazon in this realm, and it's quite likely they'll take the chance in more ways than one.