Search

The main protagonist in Kazuo Ishiguro's latest novel is Klara, an artificial friend whose existential goal is to be a child's companion. Some aspects of this fictional narrative have begun to gradually enter our daily lives. Products reminiscent of Klara are available abundantly on the market: smart toys, adaptive learning applications, and companion robots. Children can relate to these products and perform activities together with them. Preliminary research has shown fundamental differences between existing technologies and these emerging children's digital companions. However, we still do not know much about their benefits and risks. This paper explores different and even contradicting perspectives on the phenomenon. We present the discussion from four perspectives -1) temporality, 2) use, 3) trust and ethics, and 4) sociotechnical design -and conclude the paper with an agenda for interdisciplinary IS research. The agenda points to the need for a psychological, medical, engineering, and temporal research community to understand this emerging sociotechnical phenomenon and design its future for the better.

Inrätta en nationell handlingsplan som säkerställer skydd, stöd och stadigvarande boende för våldsutsatta oavsett kommun, föreslår forskare i ett projekt om kvinnor i hemlöshet.

The recently proposed theory of data network effects aims to account for how user value is created from the use of machine learning technology. The theory recognises the unique learning ability of machine learning, which uses large data sets for predictions and decision making. This paper offers a critical assessment of that theory, unearthing some of its strengths and limitations. The latter are transformed into a set of interrelated research questions that jointly constitute the proposed research programme into how to use machine learning technology in order to create and capture value. The paper contributes to the literature with an articulation of some novel knowledge gaps.

Artificial Intelligence (AI) has been argued to offer a myriad of improvements in how we work and live. The notion of AI comprises a wide-ranging set of technologies that allow individuals and organizations to integrate and analyze data and use that insight to improve or automate decision-making. While most attention has been placed on the positive aspects companies realize by the adoption by the adoption and use of AI, there is a growing concern around the negative and unintended consequences of such technologies. In this special issue we have made a call for research papers that help us explore the dark side of AI use. By adopting a dark side lens, we aimed to expand our understanding of how AI should be implemented in practice, and how to minimize or avoid negative outcomes. In this editorial, we build on the notion of responsible AI, to highlight the different ways in which AI can potentially produce unintended consequences, as well as to suggest alternative paths future IS research can follow to improve our knowledge about how to mitigate such occurrences. We further expand on dark side theorizing in order to uncover hidden assumptions of current literature as well as to propose other prominent themes that can guide future IS research on AI adoption and use.

Many would argue that artificial intelligence (AI) is not only technology but also a paradigmatic shift in the relationship between humans and machines. Much literature assumes that AI-powered practices substantially differ from and profoundly change organizational structures, communication, affordances, and ecosystems. However, AI research remains fragmented and often lacks clarity. While the information systems (IS) discipline can play a pivotal role in AI’s emergence and use, the discipline needs a clear direction that specifies how it can contribute and its key research themes and questions. This paper draws on a professional development workshop that we organized at the 2020 International Conference on Information Systems and the discussions that followed. We summarize and synthesize how AI has impacted organizational practices over five decades and provide views from various perspectives. We identify weaknesses in the current AI literature as measured against conceptual clarity, theoretical glue, cumulative tradition, parsimony, and applicability. We also identify direct actions that the IS research community can undertake to address these issues. Finally, we propose a next-step research agenda to guide AI research in the coming years.

Many would argue that artificial intelligence (AI) is not just a technology but represents a paradigmatic shift in the relationship between humans and machines. Much of the literature assumes that AI-powered practices are substantially different and profoundly changes organizational structures, communication, affordances, and ecosystems. However, research in AI is often fragmented and lacks clarity. While the Information Systems (IS) field can play a pivotal role in the emergence and use of AI, there is a need for a clear direction that specifies how IS can contribute and what are to be the key research themes and questions. This paper draws on a PDW at ICIS 2020 and the discussions that followed. It summarizes and synthesizes five decades of the impact of AI on organizational practices, providing views from various perspectives. It identifies weaknesses in the current AI literature as measured against conceptual clarity, theoretical glue, cumulative tradition, parsimony, and applicability. The paper concludes by identifying direct actions that the IS research community can undertake to address these issues. The final contribution is a next-step research agenda to guide AI research in the coming years.

A firm’s business model accounts for direct and indirect network effects, where the network size is a key enabler of value creation and appropriation. Additional conception of a business network’s contribution is provided by a recent advancement of the theory of data network effects, where machine learning is used to analyze large data sets to learn, predict, and improve. The more learning there is, the more value is generated, producing ever more data and learning and creating a virtuous circle. For the first time, this study combines the theory of data network effects with business model theory. The contribution lies in extending a business model’s lock-in effects through direct and indirect network effects to encompass data network effects. This paper provides a case study that supports the theoretical advancement and illustrates how this form of machine learning can increase profitability while reducing negative ecological impacts in an industrial context.

This article sets out to explore the phenomenon of willing digital disconnect by reconsidering and reworking some of the central ideas that currently fall under the umbrella of technological non-use. The presupposition of binary divisions between the dichotomies 'users'-'non-users' and 'analogue'-'digital' is put into question as the article explores the taking up of predigital technologies and the explicit and implicit disengagement from contemporary digital technologies. In short, this article asks: What does the contemporary revival of analogue technologies reveal about the social and material processes that constitute 'use', and what are the implications for the conceptual division of the terms analogue and digital? To answer these questions, the article draws on assemblage theory to describe the material and expressive performativity of social structure - that is, how individuals interact with technology. Empirical evidence comes from three illustrative cases where predigital technologies have replaced an existing digital alternative. Results emphasize the importance of understanding the material and expressive reconfigurations that underline technological use in a post-digital society in order to move beyond binary concepts such as analogue/digital or use/non-use as well as concepts such as the digital divide.

Evaluation is an essential part of design science research - a means to demonstrate qualities of artefacts and knowledge abstractions. The utilityfitness model suggests that evaluation needs to move from ‘usefulness’ measures to utility functions that incorporate the long-term evolution and survival of an artefact in its design landscape. In this paper, we interpret a process of innovative change taking place within a design system, in order to provide an empirical account of utility-fitness. We propose that the utility-fitness model pays too little explicit attention to technological-ecological fit, accountability, and robustness: Three ideals that were prevalent in the scrutinized empirical setting.

Open Innovation Frameworks

All embracing Internet Technology (IT) provides opportunities for contemporary organizations but it also presents a number of organizational challenges. For one thing, the rapid IT development have caused or forced individual organizational units to develop their own IT solutions, often resulting in a remarkable number of different solutions in one single organization. Today, this is seldom regarded as being efficient and management often works to consolidate IT. Such work requires that the IT unit’s co-workers are familiar with all the IT solutions currently present in the organization. Another challenge is that there often is a discrepancy between the provided IT solution and the need of the end user. To avoid this entails that the IT co-workers are well aware of the business needs. Various models and methods have been applied to deal with these challenges. Forming the IT unit in a matrix organization is one attempt. Enterprise Architecture (EA), with the main purpose of providing a holistic view of the organization, is another attempt which often embraces the matrix structure. The focus in this paper is to explore how an IT unit, working according to these premises, is being communicatively constituted. The analysis rests on McPhee and Zaug’s (2000) four-flows framework. Considering the prevailing conditions for the IT unit, activity coordination appears to be the most important flow in the constitution of the IT unit, which would premiere Taylor’s (2009) view of the four-flows framework. However, the results from our case study indicate that that the interdependence of the four communications flows is indispensible in order to communicatively constitute the IT unit.

Online communities have flourished in organizations in recent years, but large numbers of them fail. A deeper understanding of how participation can be promoted in online idea communities (OIC) is essential, because the most common reason of failure is low levels of participation. In this paper, we investigated how participation could be encouraged in an intra–organizational OIC. Our case organization was Sollentuna municipality in Sweden. They decided to introduce an OIC to collect ideas from staff as input when developing the municipality’s forthcoming IT strategy. We used the theory of online identity–based communities (Ren, et al., 2007) as a lens and a mixed research method comprising interviews and an analysis of the content of the OIC. While this theoretical perspective certainly has merits on its own, the study topic at hand made it necessary to complement the theory with more specific design principles that take the unique characteristics of intra–organiztional OIC into account. These included that managers were expected by staff to be core members of the community, frequent and complementing promotion activities were necessary in order for the OIC to be used, very low entry barriers were expected, and the employees expected the discussion to be focused, both in terms of content and time. The design principles need to be tested and further developed, by conducting studies on other OIC.

One of the main challenges that faces the AI-community is to express close approximations of human reasoning as computational formalizations of argument. In this paper we present a full implementation and accompanying software for defeasible adversarial argumentation. The work is based on the metalogic framework of defeasible adversarial argumentation games of [9]. The software we developed consists of: a meta-interpreter, a declarative implementation of the argumentation game model and a graphical interface developed in Java that shows the results of the game execution and the construction of the argumentation derivation tree.

Human argumentation in general and legal dispute in particular can be seen as highly dynamic and non-monotonic to its nature. To us this suggests that logical analysis of legal argumentation needs to be conducted in a dynamical and flexible setting in which the interaction is influenced by the parties' previous arguments. To express such approximations of legal reasoning as computational formalizations of argument, applications require dealing with knowledge representations, non-monotonic logics and a game-model able to capture the interaction as a debate between two or more disputing parties. In this paper we present some intuitions regarding the features of a full implementation and accompanying software for defeasible adversarial legal argumentation.

In press

In press

In this paper we show how to capture dialogue games in Defeasible Logic. We argue that Defeasible Logic is a natural candidate and general representation formalism to capture dialogue games even with requirements more complex than existing formalisms for this kind of games. We parse the dialogue into defeasible rules with time of the dialogue as time of the rule. As the dialogue evolves we allow an agent to upgrade the strength of unchallenged rules. The proof procedures of [1] are used to determine the winner of a dialogue game.

Automated decision-making is a significant concern for theAI community and especially for multi-agent systems. Although it has long been known among scholars of rhetoric that human decision-making can be systematically influenced by skillful argumentation, there seems to be a lack of formalizations which handle the impact rhetoric has on the concealment of logical fallacies to the human mind. In this paper, we highlight the need of metacognition for the successful formal representation and interpretation of human argumentation and thus successful automated decision-making. The relevance of such investigations is illustrated with a real-world example taken from the discourse of neuroscience.