Hani Safadi, Associate Professor, Dept. of Management Information Systems, Terry College of Business / University of Georgia.

Collective Sensemaking in Online Communities. Online communities are places where people come together to discuss shared interests. They have become more important as they provide a new way to share information, coordinate action, and create knowledge. Some online communities manage to form strong bonds among members and become true communities, while others fail to find common ground and become fragmented. In this study, we examine the factors that lead to convergence or fragmentation in online communities. We take a discursive approach and rely on the emergence of unified discourse as an indicator of community convergence, and the divergence of discourse into competing topics of interest as an indicator of community fragmentation. We situate our theoretical development within the theory of sensemaking, and hypothesize factors that shape the differing outcome of collective sensemaking in an online community. We test our hypotheses using data from fifteen online communities over more than a decade. We find that community exchanges shape the outcomes of collective sensemaking. Namely, community reciprocity, clustering, and density enable collective sensemaking whereas community transitivity and preferential attachment inhibit it. Our findings contribute a nuanced understanding of how online communities evolve and contribute to the scholarship of collective sensemaking in distributed environments. The study adds to scholarship on online communities by explaining outcomes of convergence and fragmentation and by highlighting the internal dynamics that explain them. It contributes to scholarship on collective sensemaking by documenting what activates and inhibits it in distributed settings where communications unfold over time.

Alexandre Renaud, Professeur associé de stratégie, EM Normandie (France).

USING BIBLIOMETRICS TO CONDUCT LITERATURE REVIEWS – Coping with scientific overload: How bibliometrics could help you to conduct a literature review in the age of big bibliographic data.

The multiplication of publications combined with the digitization of scientific journals constitutes a large scientific corpus immediately accessible and promotes knowledge sharing and the acceleration of scientific production. Therefore, how to cope with this profusion in what Jung (2015) calls the era of “big bibliographic data” and deal with the risk of information overload that affects the entire scientific and student community?

Bibliometrics, a cross-disciplinary science that “studies bibliographic material in a quantitative way” (La Paz et al., 2020), can help review and evaluate a large corpus of scientific literature: from simple support in research projects to more advanced sociology of science that seeks to understand the intellectual roots of the emergence of concepts, themes or scientific disciplines.

The seminar will begin with an introduction to the origins of bibliometrics, followed by a presentation of two relevant techniques for studying the content of bodies of literature: co-citation analysis and bibliographic coupling analysis. Then, we will introduce some application examples from published works and different technical solutions to conduct such work. Finally, we will take time to discuss the method, its potential in your research, and the usual issues faced during the publication process.

Marco Marabelli, Associate Professor of Information and Process Management, Bentley University.

Machine Binary: The Politics of Algorithms and DEI Implications. In this presentation, Professor Marabelli first introduced his forthcoming book titled “Machine Binary: The Politics of Algorithms and DEI Implications” (Note: As per the contract with Palgrave MacMillan, the book is due in June 2023, so most of it is still a work in progress). In this book, he has drawn on research on the political role of algorithms that he has conducted for the past several years with several colleagues and friends. Most notably, Professor Marabelli has drawn on his work with Professor M. Lynne Markus on IT ethics digital privacy and academic misconduct and ethics; and on his work with Sue Newell on opportunities and challenges of algorithmic decision making, politics of automated systems and the metaverse. In the second part of his presentation, He has specifically delved into Ch. 6 and 7 of his forthcoming book, which concern the role of institutions in regulating (or not) algorithms in various countries and geographical areas (mainly North America, Europe and China). He has contextualized the crucial role of institutions with respect to the development of the metaverse.

Shamel Addas, Associate Professor and Distinguished Faculty Fellow of Digital Technology, Smith School of Business / Queen’s University.

User Interactions with Intelligent Conversational Agents: Are Looks Everything? Extant studies on conversational agents examine the impact of anthropomorphic design elements, so-called social cues, on users’ perceptions and behaviors. However, the empirical results are mixed, and there is no consensus regarding the impact of anthropomorphized conversational agents on user outcomes. In the present study, we examine how the degree of anthropomorphism (appearance and social capabilities and the interplay between them) of an intelligent, embodied conversational agent is related to users’ perceptions, intentions, and behaviors. To do so, we draw on the Computers are Social Actors and Uncanny Valley theories, as well as the principles of agent design, to create a conversational agent that provides financial advice with natural language understanding and text-to-speech capabilities. In a controlled 3×2 experiment with 280 users, we varied the appearance of the embodied agent (robotic-like, caricature, human-like avatar) and its capabilities (social, non-social). We found that anthropomorphism features had both unfavorable effects (higher appearance anthropomorphism increased uncanniness) and favorable effects (higher social capability anthropomorphism increased social presence), but the fit between appearance and social capability lowered uncanniness; this increased trust both directly and indirectly through social presence. We also found that when participants had higher trust in the agents, they had higher usage intentions, engaged in more self-disclosure, and had higher likelihoods to accept the agents’ recommendations. Our study contributes to the literature by providing insights into the different paths that link the social cues of conversational agents to user perceptions, intentions, and behaviors. It also provides important contributions for design and practice.

Atelelier hybride sur l’apprentissage-machine / Hybrid workshop on Machine Learning

Applications of ML in scientific research. Considering MISQ’s recent editorial on machine learning in IS research (MIS Quarterly Vol. 46 No. 1 pp. iii-xix / March 2022), this workshop is designed to help us understand possible applications of machine learning in the study of IS phenomena. For example, a classification problem will be presented including its solution, which predicts if a customer will subscribe to a new product that a Bank is marketing (based on various parameters like age, education, loan, campaign, etc – mix of categorical and numeric data).

Présentateur : Reece Teramoto is an Application Engineer at MathWorks. He studied computer science and electrical engineering at the University of Portland. In 2017, he joined MathWorks and is currently an AI specialist supporting aerospace and defense accounts such as NASA and the US Army.

Jason Benneth Thatcher, Professor, Dept. of Management Information, Fox School of Business / Temple University.

Defining Individual Digital Resilience: Conceptualization, Measurement, and a Configurational Perspective. Extending work that suggests resilience drives organizations and systems’ capability to navigate environmental change, we investigate how digital resilience helps users shield against and recover from adversity caused by a jarring event that impacts the use of IT. We define digital resilience as the ability of users to shield against and recover from the adversity caused by a jarring event that impacts the use of IT. Using evidence from three studies, we develop a midrange theory of digital resilience, which (1) defines digital resilience and its dimensions, (2) offers a typology of digital resilience, and (3) connects digital resilience to the context for technology use. Study 1 uses quantitative methods to glean insight from three surveys (a q-sort, factor analysis, and test of predictive validity) to conceptualize and operationalize digital resilience as being comprised of the two dimensions of shielding and recovery. Study 2 employs qualitative methods to elicit seven factors contributing to digital resilience based on interviews with 14 teachers who navigated a jarring event in a rapidly digitizing workplace. Study 3 applies QCA analysis to data from 156 individuals who worked for an organization when a jarring event occurred in order to glean insight into how configurations of the identified factors contribute to resilience (e.g., high or low shielding and high or low recovery). We develop a nuanced context-specific understanding of how digital resilience empowers employees to perform in a digitized workplace and enables them to navigate rapidly changing technology and digital work contexts.

Jan Christof Recker, Nucleus Professor for Information Systems and Digital Innovation, University of Hamburg (Germany).

iRepair or I Repair? Dynamics of Control Enactment in Hybrid Product Platform Ecosystems. Not all products at the core of platform ecosystems are based on software alone. Many products, such as wearables, mobile phones, video game consoles and others, are hybrid, meaning that they feature both digital components, such as algorithms and data, and physical components, such as electronics, casing, and mechanical parts. We study how control is enacted in platform ecosystems with such hybrid products at their core between original equipment manufacturers and licensed platform partners on the one hand, and independent third-party complementors on the other hand. Through a longitudinal study of repair service provision in the platform ecosystem around the Apple iPhone, we examine how over a period of thirteen years the different actors used not only different modes (such as input, process, and output) but also different means (physical, contractual, and digital) for enacting control over each other’s actions. We show how the increasing use of digital means for enacting input, process, and output control have changed the nature of the dialectics between platform owner and independent complementors, and how digital means of control enactment increasingly allow platform owners to dominate and reinforce control over hybrid product platform ecosystems. We detail the implications our study makes to the literature, and we discuss the practical implications our study provides for the regulation of such ecosystems.

Théophile Demazure, Candidat au doctorat, Tech3Lab / Dép. de TI (HEC Montréal).

Enhancing Sustained Attention: A Pilot Study on the Integration of a Brain-Computer Interface with an Enterprise Information System (par T. Demazure, A. Karran, P.M. Léger, É. Labonté-LeMoyne, S. Sénécal, M. Fredette, & G. Babin (2021). Article publié dans le Business & Information Systems Engineering, 63(6), 653-668.

Arguably, automation is fast transforming many enterprise business processes, transforming operational jobs into monitoring tasks. Consequently, the ability to sustain attention during extended periods of monitoring is becoming a critical skill. This manuscript presents a Brain-Computer Interface (BCI) prototype which seeks to combat decrements in sustained attention during monitoring tasks within an enterprise system. A brain-computer interface is a system which uses physiological signals output by the user as an input. The goal is to better understand human responses while performing tasks involving decision and monitoring cycles, finding ways to improve performance and decrease on-task error. Decision readiness and the ability to synthesize complex and abundant information in a brief period during critical events has never been more important. Closed-loop control and motivational control theory were synthesized to provide the basis from which a framework for a prototype was developed to demonstrate the feasibility and value of a BCI in critical enterprise activities. In this pilot study, the BCI was implemented and evaluated through laboratory experimentation using an ecologically valid task. The results show that the technological artifact allowed users to regulate sustained attention positively while performing the task. Levels of sustained attention were shown to be higher in the conditions assisted by the BCI. Furthermore, this increased cognitive response seems to be related to increased on-task action and a small reduction in on-task errors. The research concludes with a discussion of the future research directions and their application in the enterprise.

Bogdan Negoita & Gregory Vial, Membres du GReSI et Professeurs agrégés, Dép. de TI (HEC Montréal).

Codebase Forking and Open-Source Software Development Project Sustainability: Evidence from GitHub. At a time when organizations increasingly depend on open-source software to enable and to support their operations, the software development projects that provide their information technology (IT) solutions are developed and maintained by communities of heterogeneous developers, with fluid memberships. Given these conditions, the sustainability of software development projects is a particularly important business issue. Building on the idea of forking as “an individual developer’s behavior of copying an existing project’s code base” to start a new, independent software development project, we ask the question: How does forking impact the sustainability of software development projects? Informed by a literature review of relevant extant literature, we develop a variance-based model that explains how different types of software development forks help enable software development project that exhibit source code use, as well as software development and maintenance activity over the long run. To test our deductively generated research model, we propose and detail an empirical approach based on the collection and analysis of longitudinal digital trace data originating from software development projects hosted on GitHub. Our expected findings relate 1) to conceptualizing the multiple types of forking exhibited during the software development process; 2) to explaining the implications of different types of forking for software development project sustainability; 3) to developing a set of best practices in relation to the use of GitHub-sourced digital trace data.

Ulrich Gnewuch, Chercheur postdoctoral, Institute of IS and Marketing, Karlsruhe Institute of Technology (Allemagne). Chercheur invité du Tech3Lab (HEC Montréal), hiver / printemps 2022.

More than a Bot? The Impact of Disclosing Human Involvement on Customer Interactions with Hybrid Service Agents. The proliferation of hybrid service agents—combinations of artificial intelligence (AI) and human employees behind a single interface—further blurs the line between humans and machines in online service encounters. While much of the current debate focuses on disclosing the non-human identity of AI-based technologies (e.g., chatbots), the question of whether or not to disclose the involvement of human employees working behind the scenes has received little attention. We address this gap by examining the impact of human involvement disclosure on customer interactions with a hybrid service agent consisting of an AI-based chatbot and human employees. Results from a randomized field experiment and a controlled online experiment show that disclosing human involvement influences how customers communicate with the hybrid service agent in such a way that they exhibit a more-human oriented communication style when a disclosure is made before or during the interaction. This effect is driven by impression management concerns that are activated when customers become aware of humans working in tandem with the chatbot. Furthermore, it leads to an increase in the workload of employees because fewer customer requests can be handled automatically by the chatbot and therefore are delegated to a human. These findings provide novel insights into how and why disclosing human involvement impacts customer communication behavior, shed light on its negative consequences for employees who work in tandem with a chatbot, and help managers understand the potential costs and benefits of providing transparency in customer–hybrid service agent interactions.

Pierre-Majorique Léger, Membre du GReSI, Codirecteur du Tech3Lab et Professeur titulaire, Dép. de TI*, & Sylvain Sénécal, Codirecteur du Tech3Lab et Professeur titulaire, Dép. de marketing* (*HEC Montréal).

La nécessité est la mère de l’invention : Mesurer à distance les réactions psychophysiologiques des utilisateurs en interaction avec une technologie. Comme tous les autres laboratoires d’expérience utilisateur (UX), le Tech3Lab de HEC Montréal a dû suspendre ses activités de recherche en présence de sujets humains au début de la pandémie. Toute l’équipe scientifique et professionnelle du Tech3Lab s’est lancée dans un effort colossal pour transférer en ligne les activités de collecte de données. La nécessité étant mère de l’invention, l’équipe du Tech3lab a développé un nouvel écosystème technologique appelé COBALT, permettant de mesurer précisément et de manière très riche les réactions psychophysiologiques des utilisateurs en interaction avec un artefact numérique et ce, à distance. Dans cette conférence, Pierre-Majorique Léger et Sylvain Sénécal reviennent sur les défis que leur équipe ont relevés durant la pandémie. Ils discuteront également des opportunités de recherche offertes par ce nouvel écosystème méthodologique en présentant des exemples de projets de recherche en cours au Tech3Lab.

Camille Grange, Membre du GReSI et Professeure agrégée*, & Félix Joly, Étudiant à la M. Sc.* (*Dép. de TI, HEC Montréal).

Societal acceptance of mobile contact tracing applications: the moderating effect of construal level.  When used by a majority of people, Contact Tracing Applications (CTAs) can be a most effective way to control the exponential spread of a virus within a population. However, the evidence offered from the deployment of several CTAs across the world during the COVID-19 pandemic indicates that adoption rates have remained low. The present study aims to extend the literature on the societal acceptance of these technologies by building on the lens of Construal Level Theory (CLT). We focus on the moderating role of people’s construal level on the relationship between two key factors: people’s privacy concerns (emerging from the use of CTAs) and their perceived societal utility of using a CTA – and the societal acceptance of these technologies. We are hoping that the results of this research will extend the emerging literature that uses CLT in the context of studying IT adoption in general, and that they will help provide specific recommendations to public health institutions about which design practices can be used to more effectively promote societal, mass acceptance of CTAs.

Stefan Tams, Directeur du GReSI et Professeur agrégé, Dép. de TI (HEC Montréal).

Cognitive Aging and the Struggle of Older Workers with Post-adoptive IT Use (Cet article fait présentement l’objet d’une décision de resoumission pour publication). Older workers use organizational IT in qualitatively different ways than their younger counterparts. In particular, they often focus on a few core features of an IT instead of exploiting the full range of features. This behavior is indicative of the significant reduction in post-adoptive IT use that occurs as the workforce ages. However, since the causes of this reduction remain unclear, managers and systems designers have difficulty addressing the problem of the underuse of technologies by the aging workforce. By means of a serial mediation model, this article argues that the reduction in post-adoptive IT use among older workers is caused by the decline of fluid intelligence that occurs with aging and by the impact of this decline on the ability of users to learn about new features and to adapt to them. To test the model, data were collected from younger and older users of Microsoft Excel. The results supported the model and indicated some important ways that managers and systems designers can help older workers use more features of workplace IT despite the decline in their fluid intelligence.