Skip to Main Content

Becoming Anti-Racist

Examining White Privilege in Information Frameworks

What is White Privilege?

  • White Privilege is being able to ask the question ‘Do we have to make everything about race?’
  • Urban Dictionary: A thing that has to be checked repeatedly when interacting with someone of another race, gender, kin, etc.
  • Google: inherent advantages possessed by a white person on the basis of their race in a society characterized by racial inequality and injustice
  • Wikipedia: the societal privilege that benefits white people over non-white people, particularly if they are otherwise under the same social, political, or economic circumstances.
  • My White Friend Asked Me on Facebook to Explain White Privilege. I Decided to Be Honest by Lori Lakin Hutcherson
  • White Privilege: Unpacking the Invisible Knapsack (PDF) by Peggy McIntosh (1988)

What does White Privilege mean in the world of information?

The same systemic racial oppression that occurs in other areas of life, occurs in information creation, dissemination, categorization, searching, and preservation. The traditions of universities and scholarly publishing have been western, globally north, English-language, upper-class, male, and white dominated. Early science societies and journals were white male educators and the white male wealthy who could dabble in scientific inquiry. Broad changes to scholarly publishing have occurred in the last 50 years, but English is still the predominant language for exchanging scholarship and the global north and wealthiest countries are dominant in many fields.

For more information on the role of Classics, race, and the construction of higher education, read the work of Dan-el Padilla Peralta at Princeton University. https://www.nytimes.com/2021/02/02/magazine/classics-greece-rome-whiteness.html 

Categorization: Library of Congress Subject Headings

The subject headings used to describe books and other materials are created by the United States Library of Congress and use a straight, white male assumption as the first lens. When women are used in headings, the assumption is that they are also white and straight. Other related headings are categorized with race and sexuality as secondary.

Without gender, race or geographic qualifications, “Astronauts” can be assumed to mean white American men in terms of library subjects. Anything labeled as 'American' can also be assumed to mean white. Other categories of 'American' are qualified, such as 'African American'.

Photo reads "Astronauts" with descriptors like "African American" and "Woman" preceding "Astronauts"

 

 

 

 

 

Official Library of Congress subject headings involving astronauts. Amanda Ros, CC BY

There are 4,065 subject terms containing “women” and only 444 containing “men.” For professions that are assumed to be female, nurse or prostitute, there are headings for Male Nurses and Female Nurses but only Male Prostitutes. There are terms qualifying the white women assumption, African American Women. See the work of Amanda Ros: https://theconversation.com/the-bias-hiding-in-your-library-111951

In the late 1970s, the subject heading “Afro-Americans” replaced “Negroes.” This was in turn replaced by “African Americans” or “Blacks” in 2000. In 2016, a group of Dartmouth College students and librarians petitioned the Library of Congress to replace the outdated and offensive subject heading 'illegal alien' with 'Noncitizens' and 'Unauthorized Immigration'. The Library of Congress planned to make the change but was stopped by an act of Congress. American Libraries magazine examined the current situation, fall 2020.

The B.C. First Nations in Canada has moved away from the Dewey Decimal classification system, that has a white colonial worldview, to one created in the 1970s, the Brian Deer Classification System, that better suits their indigenous peoples worldview.

Preservation: Whose Stories are in Archives?

"Archival work requires an ethics of care for the deeply personal and the deeply political… Record creation, keeping, obstruction, or misrepresentation are all acts of identity and power. Who gets to be remembered and historicized by way of record creation? Who is forgotten or purposefully silenced in history by way of omission or destruction of records? How are records themselves...used to communicate misguided notions of holistic representation, truthfulness, neutrality, and objectivity?" - Elvia Arroyo-Ramirez, 2017, Bias, Perception, and Archival Praxis

"What does it mean to be omitted from history textbooks? What are the implications of not being able to find any (or very few) traces of the past left by people who look like you, share your cultural background, or speak the same native tongue? What impact do these archival absences have on how you might understand your place in society?" - Michelle Caswell, 2014, Seeing Yourself in History: Community Archives and the Fight Against Symbolic Annihilation

Archivist Sam Winn argues that archivists are not yet equipped to deal with these questions, cross-cultural awareness is not taught in their professional education, there is hubristic trust in objectivity or neutrality, this plays to the hegemonic default - ‘white, male, heterosexual, gender-normative, upper-middle class, Global North’. ‘The default is rarely described’. Those who defy these norms are labeled radical, partisan, or activist.

"If we accept the historical fact that African Americans were at the center of American progress from the very beginning, it begs the question then, why is the historical record filled with so many silences, distortions, and erasures around Black people’s lives?" - Bergis Jules, 2016, Confronting Our Failure of Care Around the Legacies of Marginalized People in the Archives

Searching: Algorithms and Bias

  • Technological redlining
    • Reinforcement of oppressive social relationships
    • Creates new modes of racial profiling
  • Mathematical formulations that drive automated decisions in algorithms and AI are made by humans that hold all types of values and biases
  • ‘Black girls’ results in Google have changed as a result of pressures, including Dr. Noble's book Algorithms of Oppression
    • But bias in searching algorithms is still revealed by searching for ‘beauty’ or ‘Latinas’ or ‘Asian girls’ in Google Image Search
    • Google is primarily an advertising company, not a public search engine, which also contributes to its bias in displaying results.

Searching: Artificial Intelligence (AI) and Bias

  • Dr. Joy Buolamwini uncovered large gender and racial bias in AI systems sold by tech companies like IBM, Microsoft, and Amazon.
    • Error rates were no more than 1% for lighter-skinned men whilst for darker-skinned women, the errors soared to 35%.
    • The current widespread acceptance of the Black Lives Matter movement has caused companies to pause access by police to these systems. Will it change these systems for the long-term?
  • Standard machine learning can acquire stereotyped biases from textual data that reflect everyday human culture.
    • As a computer teaches itself English, it becomes prejudiced against black Americans and women.
    • Names associated with being European American were significantly more easily associated with pleasant than unpleasant terms, compared to some African American names.
  • There is work to examine the social implications of AI. See the AI Now Institute.
  • Google plans to launch AI ethics service by the end of 2020. Google, Facebook, and Microsoft have released technical tools that can help developers check for reliability and fairness. See Wired.