Graduation Year
2020
Document Type
Thesis
Degree
M.A.
Degree Name
Master of Arts (M.A.)
Degree Granting Department
English
Major Professor
Carl Herndl, Ph.D.
Committee Member
Kristen Allukian, Ph.D.
Committee Member
John Skvoretz, Ph.D.
Keywords
digital technology, rhetoric, science and technology policy, socio-technical opacity, socio-reduction, disinformation architecture, technical code, ethical design implementation
Abstract
New innovations in information management and communication technologies have produced technological assemblages which have radically altered the way people socialize and interact with the world. The most significant and ubiquitous of these technologies is what is colloquially referred to as ‘machine learning.’ Like most, if not all, technologies, machine learning models are neither wholly good nor bad. Their functional ethics are largely determined by the context in which they are employed. However, their ubiquity demands that we develop a heightened social consciousness of the way machine learning simultaneous constrains, manipulates and democratizes social processes. In order to develop better social understanding of technologies that incorporate machine learning, we must clarify how and why corporate engineers and executives scale and implement machine learning into their respective applications and services. Unfortunately, high-level calculus and computer science obscure this situation and make formulating a critical space for humanist theorists and Science and Technology policymakers an exhaustive discursive endeavor. The absence of a well understood discourse on the manner in which machine-learning algorithms are implemented represents a kind of socio-technical opacity, which obscures technological processes for contextualized corporate, design and user-motivated ethics. In order to address this problem, I propose to analyze the primary machine-learning algorithm models which organize and rank the information presented on social media newsfeed. An analysis that clarifies the function of machine learning algorithms can promote academic research and provide the impetus for Science and Technology policy incentives. Finally, this sort of analysis suggests the need for a regulatory agency for machine learning algorithms prior to their implementation into public production site environments (i.e. social media)
Scholar Commons Citation
Miller, Andrew R., "Instrumentalization Theory: An Analytical Heuristic for a Heightened Social Awareness of Machine Learning Algorithms in Social Media" (2020). USF Tampa Graduate Theses and Dissertations.
https://digitalcommons.usf.edu/etd/8971
