Labor, classification and productions of culture on Netflix

Abstract

PurposeThis paper examines promotional practices Netflix employs via Twitter and its automated recommendation system in order to deepen our understanding of how streaming services contribute to sociotechnical inequities under capitalism.Design/methodology/approachTweets from two Netflix Twitter accounts as well as material features of Netflix's recommendation system were qualitatively analyzed using inductive analysis and the constant comparative method in order to explore dimensions of Netflix's promotional practices.FindingsTwitter accounts and the recommendation system profit off people's labor to promote content, and such labor allows Netflix to create and refine classification practices wherein both people and content are categorized in inequitable ways. Labor and classification feed into Netflix's production of culture via appropriation on Twitter and algorithmic decision-making within both the recommendation system and broader AI-driven production practices.Social implicationsAssemblages that include algorithmic recommendation systems are imbued with structural inequities and therefore unable to be fixed by merely diversifying cultural industries or retooling algorithms on streaming platforms. It is necessary to understand systemic injustices within these systems so that we may imagine and enact just alternatives.Originality/valueFindings demonstrate that via surveillance tactics that exploit people's labor for promotional gains, enforce normative classification schemes, and culminate in normative cultural productions, Netflix engenders practices that regulate bodies and culture in ways that exemplify interconnections between people, machines, and social institutions. These interconnections further reflect and result in material inequities that crystalize within sociotechnical processes.

Do Men Have No Need for “Feminist” Artificial Intelligence? Agentic and Gendered Voice Assistants in the Light of Basic Psychological Needs

“…In professional contexts, for example, doctors now use intelligent image analysis tools for diagnoses ( Davenport and Kalakota, 2019 ; Kaplan et al, 2021 ) and HR managers at companies let algorithms preselect who should be invited to a job interview ( Liem et al, 2018 ; Houser, 2019 ). In personal life, data-driven AI systems recommend movies according to the user’s preferences ( Lawrence, 2015 ; Floegel, 2020 ), monitor sleeping patterns ( Alqassim et al, 2012 ; Lee and Finkelstein, 2015 ; Kolla et al, 2016 ), allow you to chat as you would do with a friend and provide emotional support (e.g., Replika, replika.ai), manage your home via smart home technologies ( Robles and Kim, 2010 ; Wilson et al, 2017 ; Gram-Hanssen and Darby, 2018 ; Marikyan et al, 2019 ), and support your every-day banking tasks ( Letheren and Dootson, 2017 ; Li et al, 2020 ). Through the use of machine intelligence, data are analyzed faster than ever before, decision processes are accelerated, monotonous tasks can be handed over to the computer, and in cases of chatbots and speech assistants, sociable connections are possible without a real human dialog partner being involved.…”

Resolving the problem of Algorithmic Dissonance: An unconventional solution for a sociotechnical problem

“…In order to minimize occurrences of Algorithmic Dissonance amongst Netflix's black users, Netflix's executives would have to explore an extensive number of imaginary scenarios that address the forms of non-knowing that its recommendation and measurement algorithms get entangled with. Such practices might attend to the pressing need of imagining and understanding systemic injustices that are embedded within these platforms (Floegel 2020).…”

Surveillance as information practice

“…Surveillance has also been framed as an appropriate subject for research within the fields of critical information studies (Carter et al, 2021) and crisis informatics (Reynolds et al, 2022). Likewise, library and information science literature has examined surveillance practices in a variety of contexts, including in the work of libraries (Fortier & Burkell, 2015; Gallagher et al, 2015; Newell & Randall, 2013a, 2013b; Randall & Newell, 2014; Tummon & McKinnon, 2018; Zimmer, 2014), in connection with labor and as a tool for behavioral control (Floegel, 2021), as a prevalent theme in LIS journals (Dewey, 2020), as embedded within commercial scholarly communication platforms (S. A. Moore, 2021), as closely linked to documentation practices (Kosciejew, 2015), as an important ethical consideration that should be taken into account when conducting research (Barriage & Hicks, 2020), and in relation to the emergence of the broader information society (Weller & Bawden, 2005). Researchers have also used surveillance records as objects of study within archival studies (Carbone, 2020).…”