Noble, Safiya Umoja (2018) “Introduction: The Power of Algorithms” Algorithms of Oppression. New YorkUniversity Press: New York.
I think the Safiya Noble text follows up nicely from the Simone Browne reading from last week. In both texts, we're made to think about the supposedly "objective" ways we gather information and form knowledge. In the Browne text, we are meant to question the objectivity of vision itself. In the Noble text, we are meant to question algorithms and the decision making processes based on those algorithms. Algorithms are not neutral, and neither is the internet.
Further, in questioning the objectivity of algorithm based decision making, Noble demonstrates the racist and sexist logics governing the ordering of information available to us online. This allows for a critical examination of the way certain bodies (specifically the bodies of Black women) are "materialized" or realized on the internet. Black women become realized/materialized on the internet according to racist and sexist logics, but those materializations are presented as objective information.
Of course, I couldn't help but pay particular attention to Noble's brief comment about glitches. She points out the exact issue with glitch or error as they are broadly conceptualized: there is a limited taking into account of the structures in which glitches or errors occur. We shouldn't understand the glitch as the sole error in the system. Rather, we should try to understand the ways in which these systems are already non-functional for certain bodies.
MacCormick, John (2012) “Search Engine Indexing: Finding Needles in the World’s Biggest Haystack”. 9Algorithms that Changed the Future: the Ingenious Ideas that Drive Today’s Computers. PrincetonUniversity Press: Princeton.
In responding to this text, I can't help but think back to last week's discussion about the value of messiness. "Messiness" or perhaps "messy data" isn't necessarily information that is free of organization. Rather, messiness might indicate that the information is being organized according to logics that are not understood by a particular user. The structure of the information might make sense to someone else - likely, the person who built the mess. In reading this piece, I can't help but feel that the variety of how people organize information (online and in their heads) isn't fully being taken into account. For example: how well does a search engine's indexing and matching work when someone's primary language isn't English? How well do these data comparisons work if the user is operating on a completely different organization of linguistic units to begin with? It's clearly not as simple as finding a word or particular set of symbols a certain number of times on a webpage. Language barriers likely come into play when thinking about the "nearness" of the terms being searched. If the user is operating on a different understanding of linguistic "nearness" that isn't compatible with English, how much sense does this search engine process actually make?
@girleffect, “Case study: Harassmap–Changing Attitudes to Harassment and Assault in Egypt.”
While this is a very cool project, I feel a bit uncomfortable with the minimal context being presented here. For example, who are the women being harassed? Are they working class? Are they queer or trans? Do they come from a specific area? I'm having trouble understanding how cellphone reports alone could address the problem of sexual harassment. If the problem is to be taken seriously, shouldn't there at least be more information about who is being attacked? But, of course, questions like that slide easily into justifications for surveilling an already vulnerable group of people. Finally, I think it's important that this project makes distinctions between state and ordinary citizen power. State and legal authority are not always invested a vulnerable person's best interests, after all…