How mining Big Data can produce fake news and voter manipulation


The opening chords of Creedence Clearwater Revival’s “Bad Moon Rising” rocked a hotel ballroom in New York City as a nattily dressed British man strode onstage several weeks before last fall’s U.S. election.

I see the bad moon rising,

I see trouble on the way

The speaker, Alexander Nix, an Eton man, was very much among his own kind—global elites with names like Buffett, Soros, Brokaw, Pickens, Petraeus and Blair. Trouble was indeed on the way for some of the attendees at the annual summit of policymakers and philanthropists whose world order was about to be wrecked by the American election. But for Nix, chief executive officer of a company working for the Trump campaign, that mayhem was a very good thing.

He didn’t mention it that day, but his company, Cambridge Analytica, had been selling its services to the Trump campaign, which was building a massive database of information on Americans. The company’s capabilities included, among other things, “psychographic profiling” of the electorate. And while Trump’s win was in no way assured on that afternoon, Nix was there to give a cocky sales pitch for his cool new product.

“It’s my privilege to speak to you today about the power of Big Data and psychographics in the electoral process,” he began. As he clicked through slides, he explained how Cambridge Analytica can appeal directly to people’s emotions, bypassing cognitive roadblocks, thanks to the oceans of data it can access on every man and woman in the country.

After describing Big Data, Nix talked about how Cambridge was mining it for political purposes, to identify “mean personality” and then segment personality types into yet more specific subgroups, using other variables, to create ever smaller groups susceptible to precisely targeted messages.

To illustrate, he walked the audience through what he called “a real-life example” taken from the company’s data on the American electorate, starting with a large anonymous group with a general set of personality types and moving down to the most specific—one man, it turned out, who was easily identifiable.

Nix started with a group of 45,000 likely Republican Iowa caucusgoers who needed a little push—what he calls a “persuasion message”—to get out and vote for Ted Cruz (who used Cambridge Analytica early in the 2016 primaries). That group’s specifics had been fished out of the data stream by an algorithm sifting the thousands of digital data points of their lives. Nix was focusing on a personality subset the company’s algorithms determined to be “very low in neuroticism, quite low in openness and slightly conscientious.”

Read the source article in Newsweek.