The Blog of Brian Ballard Quass

Brian's response to Unbiased Computer Confirms Media Bias by Rachel Ehrenberg in Science News, April 17, 2015.

Unbiased Bias




Interesting and well-written article. But it generates three concerns/observations for me:

1) The term "unbiased computer" can be dangerous in the wrong hands. Google will tell you that its search engine algorithms are unbiased -- and, indeed, Google will imply that this is so BECAUSE the algorithms are computerized -- but the algorithms are always written by human beings, and if a philosopher were invited in on the process of writing those algorithms, they could point out numerous debatable assumptions (or biases) that are inherent in their formulation (as in Google's case, the number of "backlinks" is assumed to be an indicator of site worthiness -- whereas truly original and groundbreaking work will often be incomprehensible at first sight to the hoi polloi, hence garner few if any backlinks without publicity of some kind, and so will languish under a search algorithm that rewards pre-established, "in-your-face" utility and worth).

Many I.T. advocates are, in my opinion, dangerously wrong about this. I recently saw Walter Isaacson describe Google search in effusive terms (at the Computer History Museum), praising the way it brings quality to the top. But a close (indeed, even a superficial) look at many Google search results shows that utility and influence (established names, sites ending in .edu) are the way to the top of Google's index -- not novelty and insight. Indeed, I often find whole pages of Google search results that all lead to the exact same (admittedly) uber-useful text or images, copied almost verbatim from website to website. In such cases, Google isn't bringing quality to the "fore"; they're merely facilitating access to the cut-and-paste info that the average surfer is looking for. There's no doubt a role for such a search engine, but we shouldn't delude ourselves into thinking that Google's one-trick-pony algorithms are the be-all-and-end-all when it comes to online search.

2) While there are no doubt insights to be gained from the software coding in question, it could also become just another tool to drive liberals and conservatives apart, by better identifying the individual Web surfer's political leanings and ensuring that those surfers are never pestered by links to material that is at odds with their existing mind set.

3) In some ways, as the article itself admits, this is just another one of those cases where science is finally catching up to common sense. Whether you read the D.C. "Afro-American" or the "Washington Times", you'll be hard-pressed to find outright lies or completely unconvincing logic in either of them. That's not where the difference lies. The difference between liberals, conservatives, and others lies in their answer to the following question: "What stories -- and incidents within those stories -- are worthy both of coverage and analysis in the first place?" The two above-mentioned publications provide radically different answers to that question, and so the public sees no overall Hegelian synthesis arising from the two positions -- but rather two sets of "preachers" preaching to the almost thoroughly segregated converted.




computer bias, media bias, rachel ehrenberg, science news













People who enjoyed this post went on to experience life-changing epiphanies upon listening to Brian's one-man-band electronic keyboard music written and performed under the nom-de-piano of Quasar Nibs -- which, check it out below!



Copyright 2014, Brian Ballard Quass, Quass.com No reprinting or reuse without written permission of Brian Ballard Quass: quass @ quass dot com.