Congress talk­ing to Cam­bridge An­a­lyt­ica to cre­ate buzz around Rahul

The Sunday Guardian - - The Big Story - CON­TIN­UED FROM P1

sponse till the time this re­port went to the press. The re­sponse of Congress’ so­cial me­dia head, Divya Span­dana, too is awaited on the mat­ter.

Sources said that if things work out with Cam­bridge An­a­lyt­ica, the Congress, which is ex­plor­ing newer ways to fight elec­tions un­der Rahul Gandhi, would use the ex­per­tise of the com­pany dur­ing the As­sem­bly elec­tions of Mad­hya Pradesh, Ra­jasthan and Ch­hat­tis­garh that are due in less than a year.

How­ever, when it comes to us­ing the ex­per­tise of Cam­bridge An­a­lyt­ica, the larger fo­cus of the party will be the 2019 gen­eral elec­tions.

The re­cent “im­age-makeover” of Rahul Gandhi on so­cial me­dia, which the Congress says is elic­it­ing favourable and pos­i­tive re­sponses, is be­ing at­trib­uted by some po­lit­i­cal ob­servers to these re­cent de­vel­op­ments.

Cam­bridge An­a­lyt­ica is not new to oper­a­tions in In­dia. In 2010, the or­gan­i­sa­tion worked with the JDU-BJP com­bine, which went on to win the Bi­har As­sem­bly elec­tions. As per its own state­ment, Cam­bridge An­a­lyt­ica was “con­tracted to un­der­take an in-depth elec­torate anal­y­sis and iden­tify the float­ing/ swing vot­ers for each of the par­ties and to mea­sure their lev­els of elec­toral ap­a­thy”. It was also tasked to or­gan­ise and cre­ate a com­mu­ni­ca­tion hi­er­ar­chy to in­crease sup­porter mo­ti­va­tion.

The BJP used so­cial me­dia with a great deal of suc­cess in the 2014 gen­eral elec­tions, some­thing which the Congress failed to do.

Cam­bridge An­a­lyt­ica, whose work­ing is be­ing scru­ti­nised by the US House Per­ma­nent Se­lect Com­mit­tee on In­tel­li­gence for al­leged in­ter­fer­ence by Rus­sia to “help” Trump in the 2016 pres­i­den­tial elec­tions, spe­cialises in what ex­perts de­scribe as “psy­cho­graphic” pro­fil­ing. This ba­si­cally means col­lect­ing and us­ing on­line data to cre­ate per­son­al­ity pro­files of vot­ers and pre­dict their re­sponses that are then used to de­velop specif­i­cal­ly­tai­lored con­tent to in­flu­ence vot­ers in sup­port of or against an “idea”.

The or­gan­i­sa­tion is be­ing in­ves­ti­gated by the US com­mit­tee for al­legedly pro­vid­ing, know­ingly or un­know­ingly, “help” to Rus­sian hack­ers in

their ef­forts to dis­trib­ute “fake news” and other forms of mis­in­for­ma­tion dur­ing the 2016 cam­paign.

As per peo­ple fa­mil­iar with the work­ing of Trump’s “dig­i­tal army”, which was taken over by Cam­bridge An­a­lyt­ica in June 2016, the dig­i­tal oper­a­tions su­per­vised by it had a “mas­sive” im­pact on the elec­tions in spread­ing pro-Trump mes­sag­ing via au­to­mated bots and tar­get­ing his com­peti­tor, Hil­lary Clin­ton, through neg­a­tive pro­pa­ganda. As per re­ports, Trump’s bots out­num­bered Clin­ton’s five to one dur­ing the cam­paign­ing.

Like other data min­ing com­pa­nies, Cam­bridge An­a­lyt­ica too is said to be float­ing mul­ti­ple vari­ants of ad­ver­tise­ments ev­ery day on so­cial me­dia. Peo­ple’s re­sponses to these are then con­tin­u­ously an­a­lysed and strate­gies evolved on the ba­sis of that re­sponse. These ads are pri­mar­ily floated through bots on so­cial me­dia plat­forms. The ads that get liked, shared, and retweeted the most, are then re­pro­duced and re­dis­tributed, based on where they were pop­u­lar and the peo­ple they ap­pealed to. Ul­ti­mately, this leads to data com­pa­nies be­ing able to de­velop psy­cho­log­i­cal pro­files of in­ter­net users.

Cam­bridge An­a­lyt­ica col­lects data from Face­book and Twit­ter and from other sources like tele­vi­sion pref­er­ences, trav­el­ling habits, shop­ping habits from other third-party or­gan­i­sa­tions and use all this in­for­ma­tion in “be­havioural mi­cro tar­get­ing” that “tar­gets” emo­tional bi­ases. Com­pa­nies like Cam­bridge An­a­lyt­ica also use real- time in­for­ma­tion to de­ter­mine which messages are res­onat­ing more with the peo­ple by ob­serv­ing tweets, retweets and Twit­ter thread dis­cus­sions that in­ad­ver­tently show the “per­son­al­ity and pref­er­ences” of in­di­vid­u­als. The model also uses “Face­book likes” and OCEAN scores to ar­rive at a con­clu­sion about a voter. OCEAN refers to a ques­tion­naire used by psy­chol­o­gists that de­scribes per­son­al­i­ties along five di­men­sions—open­ness to ex­pe­ri­ence, con­sci­en­tious­ness, ex­traver­sion, agree­able­ness, and neu­roti­cism.

In a speech last year, Alexan­der Nix, CEO of Cam­bridge An­a­lyt­ica, said, “We’ve rolled out a long-form quan­ti­ta­tive in­stru­ment to probe the un­der­ly­ing traits that in­form per­son­al­ity. If you know the per­son­al­ity of the peo­ple you’re tar­get­ing, you can nu­ance your mes­sag­ing to res­onate more ef­fec­tively with those key groups.”

The model de­vel­oped by Cam­bridge An­a­lyt­ica uses on­line data to iden­tify the per­son­al­ity pro­files of ev­ery Amer­i­can adult. Nix had once claimed to have “some­where close to 4,000-5,000 data points on ev­ery adult in the US”.

Their mod­els are based on the psy­cho­me­t­ric re­search de­vel­oped by Michal Kosin­ski, who was a PhD can­di­date at the Univer­sity of Cam­bridge ( hence the name “Cam­bridge An­a­lyt­ica”). Koin­ski il­lus­trated that merely on the ba­sis of analysing 10 Face­book “likes”, by us­ing his model, he was able to eval­u­ate a sub­ject bet­ter than his or her av­er­age work col­league, “Sev­enty likes” were enough to outdo what a per­son’s friends knew, 150 what their par­ents knew, and 300 “likes” what their part­ner knew.

Newspapers in English

Newspapers from India

© PressReader. All rights reserved.