3 Matching Annotations
  1. Nov 2018
    1. It encourages people to associate with those who share their views, creating filter bubbles and self-reinforcing feedback loops. Vaidhyanathan argues that by training its users to elevate feelings of agreement and belonging over truth, Facebook has created a gigantic “forum for tribalism.”

      Where is the source? Thus Vaidhyanathan provides figures of filter bubbles which decrease consuming information form others perspectives? Recent research also point towards a more diverse news consumptions due to recommendation machines.

      For an overview of empirical research see: Beyond the filter bubble: concepts, myths, evidence and issues for future debates Dr. Judith Moeller, Prof. Dr. Natali Helberger, University of Amsterdam

    2. This tool allowed the Trump campaign to upload Republican voter lists, match them with Facebook’s user database, and micro-target so-called dark posts to groups of as few as twenty people.

      Anyone has a source for this? Why would they even want to target groups as small as 20 persons? That would be highly inefficient campaigning, if you can also send groups of thousand or hundreds of thousands of people a specially tailored message?

  2. Jun 2017
    1. This is one of the smartest computer scientists in the world. He is not going to splash $15m on bullshit.”

      Cadwalladr starts with a bold statement: "a Hijacked Democracy." But does Micro Targeting actually work? One of the most important questions in the CA-debate. This article does not provide new facts about the impact of CA's "special sauce".

      Arguments given:

      Argument 1 by David, ex-CA:

      He is a smart guy.

      Ok, but why is the smart guy confinced?

      Argument 2 by Tamsin Shaw:

      “The capacity for this science to be used to manipulate emotions is very well established." The arguments are not given in this article. See note from aaronslodounik below for source.

      I find this a more convincing arguments about the impact here: https://civichall.org/civicist/will-the-real-psychometric-targeters-please-stand-up/

      Also Sue Halpern notes a similar overestimation of the impact of CA's "sercret sause" in this article How He Used Facebook to Win:

      After the initial alarm that an obscure data firm might have wormed its way into the American psyche deeply enough to deliver the election to Trump, critics began to question what Alexander Nix, the head of Cambridge Analytica, called the firm’s “secret sauce,” the algorithms it used to predict a voter’s psychological profile, what is known as “psychographics.” Confessore and Hakim’s article about the firm, which appeared on the front page of the Times, quoted numerous consultants, working for both parties, who were dismissive of the firm’s claims. The mathematician Cathy O’Neil, in a commentary for Bloomberg, called Cambridge Analytica’s secret sauce “just more ketchup.” Using psychological traits to craft appeals to voters, she wrote, wasn’t anything new—every candidate was doing it.

      Ealiers in 2012 Ethan Roeder (leader of one of the most sopisticated, data-driven campaigns in U.S history) writes in an op-ed in The New York Times:

      How do we predict wheter people are going to vote or not? We look at the voter file. It tells us how often a person votes, althought not for whome. Not all strategists agree about how to interpret this information, but the source of the data is no sectret.

      He articulates limits in general, and that it is limited specifically to information contained in public records. (More in Hacking the Electorate by Eitan D. Hersh p. 12)

      So I wonder if the CA team has so much more to manipulte with in their big database?