Conference Report: Algorithmic Cultures

Originally published in Section Culture: Newsletter of the ASA Culture Section. Fall 2019. Vol. 31 Issue 2.

114th American Sociological Association Meeting,
New York, NY Aug. 10-13, 2019

Algorithmic Cultures – Panel at ASA2019

Dan M. Kotliar
The Hebrew Univ. of Jerusalem & Stanford University

In a paper session organized and presided by Angèle Christin (Stanford University), panelists presented their work on the interplay between algorithmic systems and cultural processes. The papers offered five empirically-based perspectives on the ways in which algorithms affect – and are being affected by – various cultural forces, and thus emphasized the important role cultural sociology can play in studying the algorithmization of social life.

Eszter Hargittai (Univ. of Zurich) presented her work with Jonathan Gruber, Jaelle Fuchs, Teodora Djukaric and Lisa Brombach (all from Univ. of Zurich) on the methodological challenges in studying people’s algorithm skills; namely, their awareness and understanding of algorithms. Their proposed methodology aims to tackle some of the main challenges in studying algorithms and especially the difficulty of studying a “black boxed” subject, about which there is no known ground truth. Hence, Hargittai presented an interview protocol that instead of testing people’s knowledge about algorithms, focuses on people’s perceptions of the algorithmic systems they know and use. With acute attention to intercultural differences, Hargittai’s interview protocol was translated into four languages other than English and administered to people from five different countries. In that, Hargittai offers a much-needed alternative to the US-centered approach prevalent in contemporary algorithm studies.

Barbara Kiviat (Stanford University) presented her work on Americans’ views on the fair use of personal information. Focusing on people’s reactions to data use by car insurance companies, Kiviat offers to go beyond the more “traditional” privacy paradigm (that predominantly focuses on the collection of personal data), and into a more nuanced perspective that centers on reactions to how companies use the collected data. Based on three nationally representative surveys, Kiviat identified four types of views on personal data use: Permissives generally possess an “all-data-are-good-data mentality”; Moderatists tend to think that data is not fair to use; Domain Behavioralists see behavioral data as fair to use especially when the data matches the domain in question (in this case – driving); and Behavioral Skeptics consistently rate behavioral data as less fair to use. Kiviat also argued that each group has a distinct socio-economic composition. That is, people’s views on the collection and use of personal data is affected by their specific social positioning.

Caitlin Petre (Rutgers University) presented her work with Brooke Erin Duffy (Cornell University) and Emily Hund (Univ. of Pennsylvania) on discourses and practices of algorithmic manipulation. Based on a textual analysis of news articles, and of user guidelines by Google, Facebook, and Instagram, Petre discussed the ways in which platforms accuse cultural producers (journalists, photographers, musicians, social media content creators, and others) of algorithmic manipulation or “system gaming”. Petre argued that the lines between what platforms deem illegitimate algorithmic manipulation and legitimate strategic action are highly ambiguous, and are continually shifting in accordance with platforms’ interests and needs. Even so, both platforms and the press describe the distinction in strongly normative terms, portraying accused system-gamers as morally deviant and dishonest. Petre and her colleagues dub this moral boundary-drawing platform paternalism and convincingly argue that such gaming accusations form an important mechanism through which platforms establish, maintain, and legitimize their institutional power.

Julia B. Ticona (Univ. of Pennsylvania) presented her work with Alexandra Mateescu and Alex Rosenblat (both of Data & Society) on gender and occupational identity in online care work platforms. Ticona argues that contemporary research on algorithms and work tends to focus on male-dominated platforms (like ride-hailing apps) and overlook women’s experiences. Ticona’s paper offers a corrective by focusing on women care workers’ use of labor platforms. Particularly, Ticona offered to see platforms as spaces of cultural meaning-making, and argued that care platforms construct care work as a commodity, and their female workers as quantifiable products. While such platforms continually signal to workers that their online visibility is a proxy for their trustworthiness, Ticona showed that this visibility can prove problematic for some: for example, women from disadvantaged backgrounds often need to choose between their safety and their online visibility. 

Dan M. Kotliar (The Hebrew Univ. of Jerusalem and Stanford University) presented his work on “choice inducing algorithms” – algorithms that are explicitly designed to affect people’s choices. Based on his ethnographic research of Israeli data analytics companies, Kotliar has shown that the functioning, logic, and even ethics of choice-inducing algorithms are deeply influenced by the epistemologies, meaning-systems, and practices of the individuals who devise and use them. Thus, while people’s choices are increasingly affected by algorithms, they in fact stem from much longer strings of choices, made by multiple agents in multiple social settings. Kotliar also argued that the omnipresence of choice-inducing algorithms makes them incessant generators of choice that actively convert people into choosers. Accordingly, such algorithms are not programmed to induce specific choices, but more generally, to (re)construct the modern need to choose.


In sum, algorithms and culture are becoming inextricably linked. Algorithms have deep and often tangible effects on almost any cultural field, and at the same time, they necessarily stem from specific socio-cultural contexts. As the timely papers in this panel have shown, cultural sociology is uniquely positioned in its ability to produce empirically-based and theoretically-informed explorations of the ties between algorithms and culture, and provide a more nuanced understanding of algorithms’ growing power.