Removing order effects from human-classified datasets: A machine learning method to improve decision making systems

2023 | Decision Support Systems | Citations: 0

Authors: Romanov, Dmitry; Molokanov, Valentin; Kazantsev, Nikolai; Jha, Ashish Kumar

Abstract: Although recent developments in Artificial Intelligence (AI) and machine learnin ... Expand

Abstract: Although recent developments in Artificial Intelligence (AI) and machine learning (ML) aim to enhance the fairness and transparency of decision-making systems, research has found that neural networks (or other similar AI techniques) are still effected by human cognitive biases due to the training datasets. In this study, we focus on order effects, i.e., when the input of information impacts human perception and the decisions resulting from this information. We propose the Order Effect Removal Method (OERM) for handling the order effect which leads to bias and for helping organizations remove these biases from their training datasets and, therefore, from auto­ mated decision-making systems. Using design science principles to theoretically create, test, and validate the method, we can eliminate the order bias even in basic classification systems. Furthermore, the method can be applied in a multidisciplinary context, where an AI-based algorithm substitutes for manual work. Collapse

Topics: decision making artificial intelligence big data cloud computing enterprise resource planning

Methods: naïve bayes design science artificial neural network classification method kernel theory

Theories: probability theory

They Work with Data and Do Some Science: How Identity Conflict Turns Data Professionals away from Data Science

2023 | International Conference on Information Systems | Citations: 0

Authors: Rosa, Marta; Aikath, Devdeep; Mayer, Anne

Abstract: Data science is widely perceived as an attractive, lucrative, and prestigious e ... Expand

Abstract: Data science is widely perceived as an attractive, lucrative, and prestigious emerging occupation. Research so far has focused on understanding data scientists' practices and identity work associated with establishing and legitimizing this new occupation. This work, however, is not sufficient to explain a phenomenon we observed whereby professionals rejected the opportunity to adopt this new occupational identity. To understand why professionals may not want to be labeled data scientists, we analyzed 43 interviews with data professionals at an educational measurement company in the U.S. Despite a clear steer from management towards the data science label, many interviewees stuck to their established professional identities. In our preliminary findings, we use the literature on identity conflict as a lens to make sense of our observations. By identifying three types of conflicts: 1) task conflict, 2) role conflict, 3) tool conflict, we begin to explain what turns professionals away from data science. Collapse

Topics: database system structured query language Tableau software programming language relational database

Methods: qualitative interview grounded theory machine learning personal interview qualitative coding

Trust and Vulnerability in the Cybersecurity Context

2023 | HICSS | Citations: 0

Authors: Rosalind Searle, Rosalind; Renaud, Karen

Abstract: Cybersecurity attacks offer a pertinent context within which to examine the curr ... Expand

Abstract: Cybersecurity attacks offer a pertinent context within which to examine the currently under-explored dynamics of trust and distrust and their consequences for organisations and their employees. Such attacks wreak havoc in organisations, costly both in financial and reputational terms. We outline the dynamic, multi-level processes following initial attack making salient vulnerability to an unseen exploiter, and then making germane trust and trustworthiness in a number of relationships in an employing organisation. Drawing from events theory, we devise a multi-level conceptual trust model of the interrelations between the emotional, cognitive and social processes an attack can produce, to distinguish two paths revealing their markedly different durations, magnitudes, and level of relational and threat consequences. We explore these as dynamic experiences of trust to elucidate vulnerability and its experience and management for the targeted individual, with key organisational actors. This leads to the formation of an anchoring event that creates enduring changes to multiple relationships within the organisation with consequences for individual, team and organisational resilience and risk. Collapse

Topics: IT security threat IT security social identity

Methods: conceptual modelling

Theories: general systems theory trust theory

Co-designing a Visual Inquiry Tool

2023 | DESRIST | Citations: 0

Authors: Roschnik, Alicia; Missonier, Stéphanie

Abstract: Grand challenges may be considered as wicked problems. These wicked problems are ... Expand

Abstract: Grand challenges may be considered as wicked problems. These wicked problems are best addressed with social and visual methods rather than linear and analytical ones. One type of management tool that allows to do so is the visual inquiry tool. Their rise has been observed and a design theory to design them has been proposed. The latter provides design principles but does not however provide concrete guidance on how to operationalise them. Meanwhile, co-design has been suggested as a way to address wicked problems by using the collective creativity of the users who are experts of their own everyday activities. This paper sets out to explore the design of visual inquiry tools with co-design practices. To do so, we present the design journey of a visual inquiry tool named the “Agile Culture Transformation Canvas” (ACTC). The ACTC aims to help teams work together on their culture in the context of an agile transformation. This paper deep-dives into three different co-design events we conducted with agile practitioners to design the ACTC. From these events, we extract insights regarding what went well, and what did not go so well. Based on this, we formulate suggestions regarding how one may operationalise the design principles from the design theory for visual inquiry tools using co-design practices. This adds to the prescriptive knowledge regarding the design of visual inquiry tools. In addition to the formulated suggestions, this paper presents the current version of the ACTC. Namely, the one resulting from the three co-design events we conducted with agile practitioners. Collapse

Topics: participatory design personality agile software development user centered design

Methods: design methodology design principle conceptual modelling design theory focus group

ICT for Good: Digital Accessibility in Local Government

2023 | Americas Conference on Information Systems | Citations: 0

Authors: Rosenbaum, Howard; Shermis, Michael

Abstract: This paper describes a case study of a local government in the United States th ... Expand

Abstract: This paper describes a case study of a local government in the United States that is in the process of developing an inclusive and accessible ICT infrastructure for the delivery of e-government resources and services to its citizens with specific attention paid to people with disabilities. Developing such an inclusive infrastructure has proven to be difficult for many municipalities. Attempts to comply with the legal requirements to improve the possibilities for civic engagement for people with disabilities have met with mixed results. The research reported here focuses on a municipality engaged in a complex, user-driven initiative intended to transition its existing ICT infrastructure to one that is more inclusive and accessible to all. KeywordsDigital inclusion, access, local government, disabilities federally funded programs from discriminating against people with disabilities" (Loftus, 2021; 4). The ADA extends this prohibition to state and local government through Title II (ADA.gov, 2023). According to Grisham, the ADA uses the following definition of "disability:" (A) a physical or mental impairment that substantially limits one or more major life activities of [an] individual; (B) a record of such an impairment; or (C) being regarded as having such an impairment."The ADA "guarantee[s] equal opportunity for people with disabilities in employment, state and local government services, public accommodations, commercial facilities, and transportation" and "prohibits disability discrimination by all public entities at the state and local level." (Loftus, 2021; 2). The burden is therefore placed on municipalities to accommodate people with disabilities, providing access to all services, programs, and activities and making available reasonable means by which people can engage in effective communication and participation in their local governments. This latter requirement is generally interpreted as applying to digital content, web sites, and other channels of digital communications in which governments routinely engage. Examples of reasonable accommodations can include live on-camera sign language interpreters, assistive listening devices, and closed captioning, Among the specific requirements that municipalities must meet include the following (ADA.gov, 2023):• Communicate with people with disabilities as effectively as you communicate with others.• Make reasonable modifications to policies, practices, and procedures where needed to make sure that a person with a disability can access the state/local government's programs, services, or activities • Provide program access by ensuring that individuals with disabilities are not excluded from programs because existing buildings or facilities are inaccessible to them. Collapse

Topics: information technology infrastructure web application ADA pandemic government system

Methods: case study survey qualitative interview

Theories: theory of affordance

Are Deep Learning-Generated Social Media Profiles Indistinguishable from Real Profiles?

2023 | HICSS | Citations: 0

Authors: Rossi, S.; Kwon, Y.; Auglend, O.; Mukkamala, R.; Rossi, M.; Thatcher, J.

Abstract: In recent years, deep learning methods have become increasingly capable of gene ... Expand

Abstract: In recent years, deep learning methods have become increasingly capable of generating near photorealistic pictures and humanlike text up to the point that humans can no longer recognize what is realand what is AI-generated. Concerningly, there is evidence that some of these methods have already been adopted to produce fake social media profiles and content. We hypothesize that these advances have made detecting generated fake social media content in the feed extremely difficult, if not impossible, for the average user of social media. This paper presents the results of an experiment where 375 participants attempted to label real and generated profiles and posts in a simulated social media feed. The results support our hypothesis and suggest that even fully-generated fake profiles with posts written by an advanced text generator are difficult for humans to identify. Collapse

Topics: social media Twitter social network armed conflict website

Methods: experiment deep learning survey descriptive statistic stochastic model

Federal Platform Ecosystems to Counter Monopolists: A Value-Based Vision for the Logistics Industry

2023 | HICSS | Citations: 0

Authors: Rotgang, Alex; Duparc, Estelle; Steffen, Barbara; Hövel, Steffen; Mayer, Julia; Stodolka, Wibke

Abstract: Today, we observe the platformization of many industries. Although most platform ... Expand

Abstract: Today, we observe the platformization of many industries. Although most platforms fail, a few monopolists have the potential to completely transform industries and their competitive dynamics. In contrast to this, European values, e.g., democracy and freedom, aim to protect today’s variety of companies and offers. This paper’s goal is to protect European values and companies’ competitiveness against potential monopolists. We suggest founding and governing federal platform ecosystems following the “swarm intelligence” principle where many small(er) organizations collaboratively fight off monopolists. While this is currently still a new and untested concept, we selected a use case to make it more tangible and adaptable. The government-funded project aims at a vision for the Logistics Broker (LB) – which is envisioned to become the center-piece of the German logistics industry’s federal platform ecosystem. To analyze the context, role, and stakeholders we conducted a workshop study and propose an agenda for future research. Collapse

Topics: digital ecosystem logistics management digital platform

Methods: design artifact design science

Blockchain as a driving force for federalism: A theory of cross-organizational task-technology fit

2023 | International Journal of Information Management | Citations: 0

Authors: Roth, Tamara; Stohr, Alexander; Amend, Julia; Fridgen, Gilbert; Rieger, Alexander

Abstract: Digital technologies play an important role for the delivery of many public serv ... Expand

Abstract: Digital technologies play an important role for the delivery of many public services. However, selecting and adopting the ‘right’ digital technologies is often challenging, especially for federally structured governments. Universal factors for successful adoption are hard to establish, and the particularities of federalism, such as the separation of competencies, complicate technology selection. Nevertheless, blockchain technology seems to flourish in these environments. Through a single-case study on the blockchain project of Germany’s Federal Office for Migration and Refugees, we unpack one essential factor for this success: the fit between (cross-) organizational task structure and technological properties. This fit earns the Federal Office’s project considerable credit and traction with stakeholders and partner authorities – not least because it supports the argument that the digitalization of federal systems is possible without ‘digital centralization’ and redistribution of competencies. Our task-technology fit analysis contributes to a better understanding of the adoption of blockchain in the public sector. It also provides the foundation for an extended task-technology fit theory for federally structured, crossorganizational contexts. Collapse

Topics: blockchain government system task structure user requirement IT supported collaboration

Methods: qualitative interview case study single case study literature study triangulation

Theories: theory of task technology fit

Envisioning Digital Transformation: Advancing Theoretical Diversity

2023 | Journal of the Association for Information Systems | Citations: 0

Authors: Rowe, Frantz; Markus, M. Lynne

Abstract: ... Expand

Abstract: Collapse

Topics: digital transformation sensemaking internet of things robotic process automation digital innovation

Methods: theory development theoretical contribution literature study business process modeling grounded theory

Theories: theory of change strategic management theory complexity theory organizational behavior theory

The maturation of digital objects in innovation projects: the role of interpersonal networks and team sharing atmosphere

2023 | European Journal of Information Systems | Citations: 0

Authors: Rowe, Frantz; Te’eni, Dov; Merminod, Valéry

Abstract: Innovation projects rely increasingly on digital objects as a means of developin ... Expand

Abstract: Innovation projects rely increasingly on digital objects as a means of developing and sharing knowledge within an organisation. The maturation of a digital object is important because only those objects that mature lead to actual products. There is, however, a catch 22. For objects to mature, they have to be shared; conversely, only mature objects tend to be shared. This interplay between the functional and social dimensions of object maturation and its role in product development has not been researched.Building on the relational view of objects, we theorise product development as an ensemble of myopic maturation processes of objects that recursively move between formal and informal networks in the organisation using sharing systems such as Product Lifecycle Management (PLM). Investigating patterns, i.e., sequences of events, in these maturation processes is a way to understand how firms overcome this catch 22 situation.We study four product innovation projects in a home appliances company, tracing the maturation and sharing processes of three types of digital objects, namely marketing specifications, technical specifications and virtual prototypes. We find that sharing – maturation patterns are affected by interpersonal networks and by team sharing atmosphere. We also find that sharing and object maturation interactions are coordinated informally through interpersonal networks and formally through PLM systems and stage-gate mechanisms. Collapse

Topics: innovation management knowledge sharing organizational memory problem solving new product development

Methods: qualitative interview theory development longitudinal case study longitudinal research triangulation

Theories: organizational theory