Racism and Representation in Information Retrieval
Happy Black History Month! (originally published February 2020) by S.E. Hackney On Thursday, January 23rd, Dr. Safiya Noble spoke to an overflowing room of students, faculty, and community members about her best-selling book Algorithms of Oppression. The thesis of the book, and of Dr. Noble’s talk, is that not only racism is actually built in to the search algorithms which we use to navigate the internet, but that the big players of the internet (Google specifically) actually profit off of that racism by tokenizing the identities of people of color. It does this by associating identity phrases such as “black girls” or “phillipina” with the sites with the most streamlined (aka profitable) SEO, which is often pornography. This is a system of classification explicitly based on the centering of the white experience and and othering of Black people and other people of color. However, as Dr. Noble spoke about in her talk, tweaking a search result or two to avoid offense doesn’t actually solve a systemic problem — one where white voices are treated as the norm, and others eventually become reduced to SEO tags to be bought and sold. This idea played out recently in Barnes & Noble’s miss guided Black History Month project, where public domain books where the race of the protagonist is not specified (determined by algorithm) have new cover art created for them, depicting the characters as People of Color. Rod Faulkner, who first brought this issue to widespread attention describes it as “literary blackface,” and points out, “Slapping illustrations of Black versions Read More