Google predicts your responses, guesses your online destinations
GOOGLE’S business model wants to ensure that as much as possible, ideally everything, stays in the company’s ecosystem. Run Chrome to navigate to Drive, open a Sheet, share it in Hangouts, and email to colleagues via Gmail – you’ve never left the Google bubble.
The more time and data you spend on Planet Google, the more data the giant gathers about you, your organization and your activities. With this information, it can monetize. In some ways, it’s an accepted exchange. Google provides incredibly useful tools for free, or very little (relatively), and in return, it gets to use data about you.
The latest weapon in Google’s arsenal is prediction. The company is beta testing in a new tool called Reply in Area 120, which is similar to Smart Reply, a familiar feature to Gmail users. However the new technology can also be used in other apps, and replies are customized, based on what Google thinks you might say in any particular instant.
There’s the obvious: if your Google Calendar knows you’re in a meeting, that’s one of the three suggestions to incoming messages. But Google is also attempting to learn your speech patterns, and will (it is claimed) be able to tailor responses to the way you might voice them, all based on the historical data the company has harvested.
The third-party uses may cover WhatsApp, Line, Twitter – in fact, more or less any typical platform for communication. Additionally, if Google runs your email service either under its own name or via a white-labeled platform, the contents of the email you receive (for instance, bulk marketing emails) may well be changed so it’s more palatable to you.
As well as the opener we now all expect to be personalized (“Hi Joe,” or, on occasion, “<random-greeting> <hail-name>,”), you can start to expect localised/localized spellings of any colour/color.
Last week, the company also filed a patent for predicting online behavior. Based on your search history and your web-based movements, as reported by various tracking cookies and code snippets, pages could be preloaded before you even know yourself which way you’re heading.
At #digitalnations2030, Tomas Izo of Google talking about role of Big Data for predictive outcomes – Google is exploring how to incorporate sophisticated policy elements such as equality of opportunity, fairness and bias in machine learning. pic.twitter.com/6KVSRYvtYw
— Stephanie Honey (@StephHoneyTrade) February 19, 2018
All of these technologies are, to a certain extent, predicated on users’ opting in (or not opting out of) of tracking and harvesting. But as Google’s tendrils spread more extensively, and its offerings become even more ubiquitous and, frankly, useful, the alternatives may seem to be fewer in number and generally speaking, just not as clever.
As organizations of all sizes have seen in recent years, ‘big data’ has emerged as a pool of information that can be mined for particular purposes. But one effect of the abundance of data is that it’s increasingly difficult to sift through what’s out there, to find the useful gems. Predictive technologies may be the key to cut to the chase, in data terms.