How did Donald Trump defy the swelled-head mediacracy to become President? From the fringe of the tech world a theory has emerged: It was artificial intelligence. No, not just regular AI, but rather a “weaponized” artificial intelligence, says Jonathan Albright, a data scientist and professor at Elon University in North Carolina.
The Trump campaign used an AI-powered system made by Cambridge Analytica, a U.K. Outfit that reportedly gets funding from Robert Mercer, a billionaire Trump supporter, and that was whispered to have Trump strategist Steve Bannon as a board member. (A spokesperson for Cambridge Analytica would not comment on Mercer and would say only that “Steve Bannon is not a board member.”) In any case, the company hoovers up mountains of consumer and lifestyle data—including what you watch on TV—in order to build a “psychographic” profile of you, the individual voter. It claims to possess profiles of 230 million Americans—each based on 4,000 to 5,000 discrete data points.
Armed with that insight, the company can fill your social feeds with ads and content designed to push your emotional buttons. The goal is to make you do something—such as vote.
While the notion of using hidden psychic levers to affect a person’s political behavior may sound skeevy and invasive to some, it’s really no different from what consumer goods marketers have done for years to sell soap. So, no surprise, that’s precisely where CA is headed now—applying its AI sauce to “brand and commercial marketing” so that clients (which CA won’t name) can “identify new customers and persuade them to engage with their brands.”
And AI, it turns out, is the Dale Carnegie of the modern age: precise, personal, and a whole lot faster than a door-to-door salesman in closing the deal. Super-brains can crunch mountains of data in milliseconds, A/B testing thousands of messages to see which ones work best on which people, adapting and evolving on the fly. “It’s a level of social engineering I’ve never seen before,” says Albright, who is quick to point out that CA is hardly the only one doing this—or things like it.
AI-powered machines are also now creating content, he says. An AI platform called T, for instance, creates “news” videos on YouTube, many related to politics. Each video contains photos and text culled from blogs and websites and is “narrated” by a computer voice. Albright claims T was churning out a fresh video every three to four minutes. (The professor says he has found at least 80,000 videos but reckons “there might be hundreds of thousands or even millions of them.”)
For marketers this sounds like a dream come true. If a “weaponized” AI sales agent can push people to get off the couch and actually vote, then certainly it can persuade people to choose Samsung instead of Apple (AAPL), or Chevy over Ford (F).
Our poor manipulatable psyches don’t stand a chance.
A version of this article appears in the April 1, 2017 issue of Coins2Day with the headline “AI’s Killer App? Duh…Marketing.”