15.5.2.7 Intelligent assistants will become more intelligent
Over the next decade, most people in developed countries will begin interacting with intelligent assistants multiple times a day. These assistants will do more than simply manage your schedule and act as an interface to search engines – though those will still be important functions.
Deep Question Answering systems, of which IBM’s Watson is a present-day example, will be integrated into these assistants. Classical ‘search’ will be replaced by ‘conversation’ where you ask a question, are given an answer, and can ask follow-up questions – just as you would do with a knowledgeable human. Note that these systems will be far more sophisticated than key word searches that return page after page of web links that may or may not contain the information you are looking for.
Currently, if you want an intelligent assistant to do something for you, you have to explicitly ask it to do so. I expect within the next few years our assistants will grow to the point where they will be able to offer up advice, suggestions, and answers to questions of interest without having to be asked. They will begin anticipating our needs, just as an excellent human assistant is able to do. In order for your assistant to offer this level of help, it will need to learn about you – your likes, dislikes, schedules, and routines. This is clearly the direction that Google is clearly headed with their Google Now product.
For example, when leaving on a recent trip I was surprised to find that Google Now knew my flight and hotel booking information, though I had not told it these things. Google Now apparently learned the details of my trip by scanning my email. Some people will find this creepy, and to be honest my first reaction was a kind of nervous shock – but once I started getting flight delay and gate change updates, and weather conditions for my destination city, I quickly got over those feelings and began to appreciate how Google Now could make travel a little less weary.
Google Now also does things like look at your movement patterns, figures out where you live and where you work and then feeds you route and travel time estimates – .updated with current traffic conditions.
The general idea is that the more these systems know about you the more helpful they can be – and not so coincidentally the more targeted the ads they can feed to you. Barring an extreme backlash over privacy concerns, expect this trend of your intelligent assistant knowing more and more about you to continue over the next decade.