Digital practices in institutionally complex contexts: Methodological challenges and practical solutions to understand the opaque and elusive 

Abstract

Using Digital Platforms to Support Educators’ Learning about the Societal Impact of Algorithms

by Stephen Adams, California State University Long Beach

This presentation discusses the application of two e-learning platforms in a university course to support learning about the societal implications of algorithms. The university course is part of a Master’s program in educational technology in the USA. The subject of algorithms and their societal impact raises complex interdisciplinary issues that can be challenging for students. It is useful to find ways to support students in interacting and exchanging ideas about the issues. The presentation discusses the use of
digital tools to support students’ discussion of and reflection regarding these topics.

The subject matter concerned topics including the opacity of algorithms (Tufekci, 2017), societal implications of social media and YouTube (Haugen, 2022; Chaslot, 2019), racial bias in algorithms (Buolamwini, 2018) and foundational issues in STS (e.g., Winner, 1980).One platform was used for journal articles and another platform was used for videos, such as documentaries, news stories, and Youtube materials. A platform designed for textual information, Perusall, was used to allow students to collaboratively highlight and comment on passages in articles in an electronic format. This helped make the activity of reading articles and considering the implications
of specific passages to be a more social process. A platform for videos, Playposit, was used to allow students to discuss videos as they watched them, attaching comments.

The presentation demonstrates examples of the applications of these tools, discusses how students made sense of the material using them, and reports on students’ feedback regarding the use of these tools in this area.) 

Dealing with AI-generated media:  young Finns’ authenticity identification regarding deepfake videos

by Yucong Lao, The University of Oulu

The advancement of artificial intelligence (AI) has reshaped people’s digital practices. As novel AI technology impels synthetic media to be more mature, the content of deepfakes looks more realistic, which leads people to “re-evaluate” their “very perception of reality”. Due to the improved quality, deepfakes might arouse issues related to privacy, fraud, and disinformation, and further cause negative impacts on democracy. Since deepfakes flourish on various social media platforms, young people might encounter deepfake videos in everyday lives. Therefore, it is vital to understand how young people identify the authenticity of media content through their digital practices.
The aim of this paper is to increase understanding on young people’s digital practices of dealing with AI-generated media. More specifically, the purpose of this study is to investigate how young Finns identify the authenticity and evaluate the media content of deepfake videos.

In-depth interview with an experiment is applied to collect empirical data for this study. 20 Finnish 9th grade pupils were recruited as the participants for the research project. Two selected deepfake videos were provided to the participants to watch. The participants were allowed to share their thoughts while watching videos. After watching the videos, they were asked to clarify the process of deepfake content identification and evaluation.

This study will offer a lens on young people’s digital practices in contemporary environment surrounding with massive media powered by generative AI. In addition, it can drive policy makers to revalue and improve governance on emerging AI-powered products in this digital era.) 

Applying Paradox Theory in Exploring Tensions during AI Implementation in Healthcare Organisations

by Zeng Li, University of Edinburgh

This study focuses on the conflicts and tensions arising from the implementation of AI technology in healthcare organisations, using a theoretical framework of paradox theory, and seeks to expand this finding from AI technologies and digital transformation. Specifically, we present a case study of a Chinese hospital that abandoned the implementation of an AI-based Cancer Diagnosis System. Through interviewing seven members, we find that AI implementation in healthcare organisations is highly complex and cannot be fully covered by the classification claims of learning, organising, performing, and belonging tension proposed in the paradox theory. To address this problem, we extend the classification of tension by introducing the concept of technologing tensions, which provides a more comprehensive perspective to study tension within organisations during the AI implementation stage. The contribution of this study enhances the understanding of the tensions within organisations during the implementation of AI, which facilitates the promotion of AI-based applications in healthcare and contributes to the development of paradox theory.  

Living with disabilities and improvising: exploring digital practices and knowledge transformation of people with disabilities in China

by Chenchen Ma, University of Amsterdam

Anthropology is facing new challenges brought by the Internet when new possibility in ethnography is one of the most prominent to be influenced. In a more conventional ethnography, anthropologists immerse themselves into the field sites, and what they do will get closer to what Clifford Geertz (1973) calls “thick description.” However, on the Internet, the boundaries of both space and time of ethnographic fieldwork have been challenged. A considerable amount of research has given examples of new multi-sited ethnography or purely online ethnography (Landzelius, 2006; Boellstorff, 2015). This research was conducted during COVID-19 and multiple methods of doing digital ethnography were experimented to explore the digital practices of people with disabilities in China. It focuses on how people with disabilities transform their daily experience into disability knowledge by using cellphilms. Inspired by the study of patients with chronic diseases (Pols, 2014; Pols and Hoogsteyns, 2016), I will try to present the capability of people with disabilities to translate parameters into practical courses of action in disability communities. I will also show how people with disabilities improvised on digital media platforms to apply and spread the knowledge (by experience) to a broader social group, and to get career achievements. To accomplish an “thick description,” I took one year’s run on my field. A method of “adaptive listening” (Winter and Lavis, 2020) will be applied, which theorizes “listening” as a multisensory act that complements observation by creating a way for researchers to observe digital practices and deal with more detailed data.

Cuts through complexities: methodological steps to map digital Nordic border practices

by Luna Secher Rasmussen, IT University of Copenhagen; with Ursula Plesner, Copenhagen Business School, Denmark; Elena Raviola, Gothenburg University, Sweden

This paper accounts for the development of a mapping strategy as a way to navigate the complexity of extensive and heterogeneous extant data (Charmaz 2006) in a preliminary research phase, where the aim was also to arrive at an understanding of a multiple object of study.  

The mapping strategy is developed in the context of a collaborative exploration of digital public surveillance practices at the Nordic borders – an institutionally and technologically complex phenomenon evolving over time. The methodological steps involved 1) inductive desk research for extant document data (Charmaz, 2006), which implied reflections regarding search words and object of study, 2) the development of a data excel sheet containing the elements of ‘Dates’, ‘Sets’ and ‘Tags’ drawn from contextual positioning (Ralph, 2014) and the concepts of categories and labels in situational analysis (Clarke, 2005) and analytics (Marres, 2020), including creating the possibility to cross reference content, and 3) the development of a python script to allow for making analytical cuts across a large, complex empirical material inspired by the methods of self-coding (Glazier, 2021) and semi-automated lexicon based analysis (Marres, 2020) and for the visualization of different ‘occurrences’.  

This mapping strategy adds to the STS tradition of mapping controversies and issues (Clarke, 2005; Venturini, 2010; Marres, 2015; Marres, 2020) through the uses of different digital methods. Besides offering a detailed mapping method, our paper fleshes out methodological challenges relating to capturing temporality, technicalities, and transparency issues in the material.

Organizers

Elena Raviola, University of Gothenburg; Vasilis Galis and Luna Rasmussen, IT University of Copenhagen