THE TAN­GLED TALE OF CAM­BRIDGE AN­A­LYT­ICA

Did a lit­tle-known com­pany use pro­fil­ing and big data to ma­nip­u­late the US pres­i­den­tial elec­tions?

HWM (Singapore) - - THINK - By Alvin Soon

“His sci­enti c work on the Ocean mod­els cau­tioned that these profling and pre­dic­tion tech­niques could be used in ways to ma­nip­u­late peo­ple.”

In De­cem­ber of 2016, an ar­ti­cle in the Ger­man Das Magazin mag­a­zine caused a stir. Re­porters Hannes Grasseg­ger and Mikael Krogerus wrote that a lit­tle-known com­pany called Cam­bridge An­a­lyt­ica had used pyscho­met­rics, big data and so­cial me­dia to help swing the US pres­i­den­tial elec­tions in fa­vor of one Don­ald J. Trump.

In a pre­sen­ta­tion at the Con­cor­dia Sum­mit, Alexan­der Nix, CEO of CA, re­vealed how the com­pany had crunched big data to in­flu­ence the elec­tions. It had bought per­sonal data from var­i­ous sources, ag­gre­gated this data and cal­cu­lated a per­son­al­ity pro­file based on five per­son­al­ity traits, also known as the Ocean Method.

Nix told the crowd that by cat­e­go­riz­ing vot­ers based on their psy­chol­ogy, the com­pany could de­sign specifc ads to tar­get their val­ues. “For a highly neu­rotic and con­sci­en­tious au­di­ence,” Nix said, he might be shown the threat of a bur­glary, to bet­ter sell him a gun. Nix told Moth­er­board that on the day of the third pres­i­den­tial de­bate, Trump’s team tested 175,000 dif­fer­ent Face­book ad vari­a­tions, to find the ones with the best in­flu­ence.

Michal Kosin­ski, a lead­ing ex­pert on psy­cho­met­rics – a data-driven branch of psy­chol­ogy – has demon­strated how much you can do with the Ocean Method, when com­bined with enough data. In 2008, Kosin­ski, then a stu­dent at the Uni­ver­sity of Cam­bridge in Eng­land, wrote a small quiz app for Face­book called MyPer­son­al­ity, based on the Ocean Method. Mil­lions of peo­ple took the quiz, which, to­gether with the sub­jects’ Face­book ac­tiv­ity, gave Kosin­ski and his team a wealth of big data to work with.

In 2012, Kosin­ski demon­strated some­thing star­tling: From a mere 68 Face­book likes, his team could re­li­ably pre­dict a per­son’s skin color, sex­ual ori­en­ta­tion, and whether they leaned Demo­crat or Repub­li­can. With 150 likes, they could pro­file a per­son bet­ter than their par­ents, with 300 likes, Kosin­ski’s pro­fil­ing tech­niques could pre­dict a per­son’s be­hav­ior bet­ter than their part­ners.

Nei­ther Kosin­ski nor Cam­bridge Uni­ver­sity are as­so­ci­ated with Cam­bridge An­a­lyt­ica, but the process ap­pears to be iden­ti­cal to the mod­els that Kosin­ski de­vel­oped and that he had warned peo­ple about. His sci­en­tific work on the Ocean mod­els cau­tioned that these profling and pre­dic­tion tech­niques could be used in ways to ma­nip­u­late peo­ple — and it looked like a com­pany had found a way to weaponize it for po­lit­i­cal gains. But had it re­ally? Fol­low­ing the buzz that arose with Das

Magazin’s re­port, the New York Times spoke with Repub­li­can con­sul­tants and for­mer Trump cam­paign aides, along with cur­rent and for­mer CA em­ploy­ees, who say the com­pany’s abil­ity to ex­ploit per­son­al­ity pro­files are “ex­ag­ger­ated.” Ac­cord­ing to the

NYT, CA ex­ec­u­tives “now con­cede that the com­pany never used psy­cho­graph­ics in the Trump cam­paign.” And for all the ap­par­ent suc­cess CA had with the Trump cam­paign, there was one in­con­ve­nient fact that they couldn’t es­cape. Be­fore Trump, CA had worked with Repub­li­can pres­i­den­tial hope­ful, Ted Cruz. Cruz even­tu­ally lost the Repub­li­can pri­mary to Trump and sus­pended his cam­paign for pres­i­dent.

Whither the truth about how pow­er­ful Cam­bridge An­a­lyt­ica’s tech­niques are, per­haps the fact that we are so able to be­lieve their claims and not just dis­miss them as sci­ence ction is the more telling story. We live in a sur­veil­lance econ­omy, where pow­er­ful com­pa­nies like Google and Face­book al­ready know more about us than we do about them, and we take on­line track­ing for granted more of­ten than not.

We be­lieve a com­pany like CA is pos­si­ble be­cause the tech­nol­ogy al­ready ap­pears to be here, or if not now, then soon. And that caps this tan­gled tale with a more wor­ri­some end­ing than any­thing.

Newspapers in English

Newspapers from Singapore

© PressReader. All rights reserved.