Google will transfer data and cloud processing to Android devices.

For years, Google I / O, the internet giant's annual conference for software developers, has highlighted the degree of profound transformation of mobile telephony. They had gone from a majority of standalone devices, even if they were connected to voice and data networks, to devices that did not really support the actual machine processing that made them so useful. The heavy load has increasingly taken place inside the megabrain called Google. It was powered by complex, opaque software and hardware that massaged and moved information, mostly collected from you and me, into large data centers. Our devices were becoming screens (and input systems) for the Big Tech universe, where true intelligence lived.

The power of this model was amazing. It was nowhere more obvious than in Google Maps, which is for me (and I bet many other people) the most essential use of my mobile phone, after conversations of all kinds.

But this week, in Mountain View, California, Google announced that it will transfer some of the most essential information to Android devices.

Not tomorrow, but soon. Google CEO Sundar Pichai announced in his keynote speech that he called it an "important milestone", made possible by seemingly endless improvements in hardware and software, which would greatly increase the phone's processing power. Important features that Google and the rest of the tech industry have worked on: voice recognition.

Until now, Pichai said that Google needed about 100 gigabytes of storage on its servers, as well as a constant network connection, so that speech recognition would work at a level sufficient for Daily use of each user. According to Mr. Pichai, the progress made allowed the company to bring back these 100 gigabytes to 500 megabytes and transfer them directly to the mobile device of the user. This eliminates the delays caused by the network and, he said, will transform the way people use their mobile devices.

The first example is Google Assistant, the voice feature of the new Android phones that allows you to tell the device what you want it to do for you and to get answers to your questions. The main demonstration of this phone race was more than impressive. We, the device users, have not seen it yet – Google says it will not happen until the next generation of phones appear later this year, and it's not going to happen. Is not clear if this feature will be added to current phones.

What is most intriguing about this transfer of power to users' devices is what it suggests. An associated I / O announcement gives us an overview of some of the possibilities. On the one hand, Google claims to work on ways to keep more personal data of users on devices and not in the company's databases.

Google calls a key element of this "federated learning" initiative and deploys it first in its Gboard mobile keyboard software. It's scary to realize that a mobile screen keyboard could easily capture and send back to someone else's memory everything you type. One of the reasons is how users use devices and improve the quality of data entry for the user. Federated learning means that whatever you type stays in your device. Only when many people enter the same kind of words and sentences does the mother ship learn that – and no one's personal data ends up on the company's servers.

This type of system, already used to some extent by Apple, offers considerable potential to protect confidentiality. It remains to be seen if the promise will survive business imperatives, but as our personal devices gain power and storage, there are fewer and fewer reasons – other than business surveillance – for businesses.

The main reason they keep a lot of our data is that we will ask them, unfortunately. The convenience of "cloud" storage, for backups and access from any device, is a part of our job. But there are other ways to back up and access our personal information without entrusting it to Google and other giant corporations. We should learn more about them and use them.

Not everything will soon be on our phones. Google Maps is a classic example of why not. If we want information on Google's real-time traffic, there is only one way to get them: from the network.

This will be the case for some other applications and services, but over time, as the devices we hold in our hands become more and more powerful, the possibilities are multiplying for new ways of using them. Personally, I look forward to having a truly functional Star Trek tricorder.

Time of the future
is a partnership of
Slate,
New America, and
University of the State of Arizona
which examines emerging technologies, public policies and society.