Pay no attention to the man holding the receiver; it’s the phone itself you’ll need to impress.

A small army of editors, programming directors, critics, censors, librarians and curators have shaped choices of entire generations. But just like factory workers of the 19th century found themselves inevitably replaced by the more efficient machines, human gatekeepers are giving way to a new breed of automated tastemakers – sophisticated software that separates the information wheat from the chaff and whose influence is growing as fast as the amount of information we produce.

That is, pretty fast.

In 2002, the amount of new information was estimated to double every three years.    Today, it is doubling in size every 18 months. IBM expects that soon the amount of information will be doubling every eleven hours.  47,000 radio stations around the world pump out 70 million hours of original radio programming; 21,000 TV stations create 31 million hours of news, dramas, and reality shows; and printing presses roll out over a million new book titles, 25,000 newspapers, 80,000 trade periodicals and 37,000 academic journals. (source) 

Finding a needle in this haystack of information requires superhuman powers. It is likely that in the near future these powers will come packaged in a slick portable device that stays awake around the clock, knows your tastes, detects your mood, programs your TV, fills your earbuds with infallibly pleasing music, and instructs you what to have for dinner and where, what books to read and what clothes to buy.

It will probably be a (really) smart phone.

To be an effective gatekeeper, the phone should be able to identify and deliver a single best piece of information that is uniquely appropriate in a given context; that is, it should give you a sensible fashion advice when you are standing in front of your wardrobe and recommend a good book when you are at your corner bookstore.  In other words, the device will need to have access to a library of knowledge, to understand what it is you are doing, and to pick information that best suits your immediate needs while keeping out the rest. 

In order to be machine-readable, information needs to be digital, and digitization of knowledge is happening today at a brisk pace.  In 2008, the world has produced 3,892,179,868,480,350,000,000 new bits of digital information.  That’s 487 exabytes (billion gigabytes); up from 161 exabytes in 2006, or “3 million times the information in all the books ever written.”   For comparison, all words ever spoken by human beings add up to only five exabytes.    By 2011, researchers calculate the number will reach 1,800 exabytes (or, more accurately, 1.8 zettabytes).  All these bits and bytes come from TVs, radios and telephones converting to digital formats, from newspapers publishing stories online and digitizing their archives, from companies racing to scan entire book libraries (Google alone has scanned seven million books to date), from electronic records of various transactions, from office workers emailing memos, and from regular people taking and uploading pictures and videos. By some estimates, 66% of all media consumed will be digital in 2010.

Second, the cell phone should be aware of its surroundings.   A company called Integrated Media Measurement uses cell phones to track consumer exposure to ads on TV, radio, computers and other media by matching the audio in the ads to its database.  Researchers at the Context Aware Cell Phone Project at MIT Media Lab are building a prototype of a cell phone that determines whether the user is at a restaurant, driving a car, or watching a movie at the theater.   Another Media Lab project focuses on creating a phone application that analyzes speakers’ tone of voice to identify their respective interest in the conversation.

Finally, the device will need have a system that identifies the most appropriate information. These systems now come in three flavors, and if you have a computer you are probably already familiar with all three.  They are spam filters that are designed to block stuff they think you don’t want, search engines that deliver what you explicitly demand, and recommendation and discovery engines that give you what they think you want but don’t know how to ask. 

The many recommendation engines in existence cover a wide range of approaches and subjects of expertise.  Popular online radio service Pandora.com runs the song you pick against a list of 400 musical attributes such as harmony and rhythm and selects songs that are similar in structure.   Jinni.com attempts to take the same approach to movie recommendations – it considers plot elements and structures, mood, and tone — while WhatToRent.com picks movies for you based on the results of a short personality quiz.  Other systems recommend shoes, books, clothes, TV shows, restaurants, and radio stations.   Many of these services are available on cell phones.

Some companies are already putting all these pieces together. Palo Alto Research Center has developed software for cell phones that helps people find things to do based on person’s location, time of day, and past interests. Geodelic’s application Sherpa for Android phones provides information about its user’s current location: show times and reviews if the person is at the theater, the menu if it’s a restaurant. 

Once enough people start relying on their phones for recommendations in their daily life or even delegate certain activities entirely — automatically programming the DVR, for example — advertisers will have to worry first about seeking favors with the new gatekeepers, and only then with making an impression on humans.  In a sense, their job will be similar to that of search engine optimization professionals who adjust websites according to how, in their opinion, Google ranks pages.

Suddenly, telemarketing will have a whole new meaning.