"They told me, 'Here's who we think your best friends are,'" by analyzing who got the longest calls, the most e-mails, the most texts, and even who got the longest texts from him, among other patterns.
The value of that kind of list is that it would help an automated system populate favorites, much the same way that Netflix suggests movies someone will like, Donovan said. In one example, Donovan said a TV today will turn on to the last channel watched, but it could be set up with a profile for every member of a family to turn on to the most-watched channel for each person.
"This is the difference between discovery and search, and [then] find," he said.
To expand on the concept, he said that with mapping tools and location data, algorithms could, for example, compare the day of the week with the city a user is in and then tell authorized members of the user's contacts list that the user may be on a vacation or business trip.
The level of detail that's possible from the AT&T lab experiments seems so personal and so invasive that Donovan admits it will initially be controversial to most people, even "creepy."
"The order of my 30 best friends that the list gave me was better than the order I gave it," he said. "It was creepy."
He explained that when he analyzed the list, it made him realize that he should have been in touch more often with a good friend who is an amateur hockey coach. Since it had been too long since they were in touch, he took the prompting from the list as an opportunity to send the friend a trash-talk text about hockey.
"This kind of conceptual stuff is going to move from creepy to spooky to mainstream" in coming years, Donovan said.
He admitted there would probably be enormous pushback over personal privacy initially. Also, he said manufacturers such as Apple that are proud of the design of their smartphones and tablets would still want everyone to own a device instead of using someone else's device at a friend's house or at work or in the car.
"As I look at the horizon, this industry is just beginning to hit," he said. "It's not just mobile and cloud where things can be detached from devices, whether it's the TV or the laptop." What Donovan called "disintegration" of the user from the device will allow a "lot of flexibility," he said, adding "we'll see socially amazing things."
Sign up for Computerworld eNewsletters.