According to Mark Cuban, "The world's first trillionaires are going to come from somebody who masters A.I. and all its derivatives, and applies it in ways we never thought of."
Research shows he may be right.
A 2014 Stanford University and University of Cambridge study shows that computer-based personality judgments are more accurate than those made by humans. (Think of it as the "Who Knows Me Better: Facebook or My Spouse?" project.)
First, researchers had over 86,000 volunteers complete the 100-item International Personality Item Pool, a questionnaire that measures the "big five" dimensions of personality: Openness, conscientiousness, extraversion, agreeableness, and neuroticism.
At the same time, the volunteers' coworkers, friends, family, and significant others described each volunteer by filling out a version of the Personality Item Pool. The researchers also obtained 'likes' from each volunteer's Facebook profile.
Then the computers went to work: A little linear regression here, a lot of statistical techniques (that only statisticians understand) there, and boom.
Turns out computers kick ass.
Here's how many 'likes' the computer model needed to predict an individual's preferences better than the following:
- Coworkers: 10 likes
- Friends: 70 likes
- Parents or siblings: 150 likes
- Spouse: 300 likes
Yep: If you've hit 'like' 300 times, algorithms -- like the ones Facebook uses -- know you better than your spouse does. (Just imagine what A.I. can do by adding in your all shares, tweets, search histories, comments, etc.)
But there are caveats.
For one thing, like-based models are "more diagnostic" of some traits than others. Take openness, a trait that's hard for other people to judge. (It's reasonably easy to observe when someone is neurotic; "openness," is tougher.)
That's where evaluating a digital footprint has a decided advantage. "As openness is largely expressed through individuals' interests, preferences, and values," the researchers write, "we argue that the digital environment provides a wealth of relevant clues presented in a highly observable way."
Or in simple terms, I may not know you're into abstract-expressionist art, meditation, and deep-dive TED Talks... but your devices certainly do.
On the other hand, algorithms were terrible at predicting "life satisfaction," a finding that also makes intuitive sense. I might 'like' a bunch of motivational quotes because I desperately want to live to a more fulfilling life... or because those quotes simply validate the satisfaction I already feel.
And then there's this: Sometimes a computer model knows you better than you know yourself. Where Facebook activities, substance use, field of study, and network size were concerned, the model out-predicted volunteer's self-reported traits and behaviors. (Just as the average person tends to underestimate how many calories they consume, and overestimate how many friends they have; we're all guilty of wishful thinking.)
Sum it all up, and as the researches write, "The ability to accurately assess psychological traits and states, using digital footprints of behavior, occupies an important milestone on the path toward more social human-computer interactions."
Or more to the business point, according to Cuban:
If you don't know A.I., you're the equivalent of somebody in 1999 saying, "I'm sure this Internet thing will be OK, but I don't give a sh*t." If you want to be relevant in business, you have to, or you will be a dinosaur very quickly.
There's going to be A.I. haves and have nots. If you're a have not, you might as well rip out all the computers in your office and throw away your phones. That's how impactful it's going to be.
Because the better you understand what people think, need, and want, the better you can meet those needs.