AI decision support systems (AI-DSS) increasingly advise human decision-makers, while the quality of individual recommendations cannot be directly verified at the time of use. In such settings, users must calibrate their trust in these systems, based on interface cues such as stated system accuracy and the system’s explanations. Large language models enable agentic reasoning: natural-language rationales that appear deliberative and case-specific. We argue that the convincingness of such reasoning is not a generic trust booster, but a conditioning cue that shapes how users translate capability claims into trust judgments. We test this in a 2 × 2 between-subjects Judge–Advisor experiment, manipulating stated accuracy (70

This study explores the impact of short delivery times on product returns in the context of online retailing. Using a large dataset from a global fashion retailer's U.S. e-commerce platform, we investigate whether fast deliveries characterized by below-average delivery times influence the likelihood of product returns. The analysis employs logistic regression to examine the relationship between delivery times and return rates, and additionally considers product characteristics and customer attributes. Our findings indicate that fast deliveries lead to a significant increase in the likelihood of returns, particularly among new customers. Insufficient post-purchase cognitive dissonance reduction may theoretically motivate this counterintuitive result as also indicated by a preliminary follow-up study reported in the online appendix. These insights challenge the prevalent assumption that the shortening of delivery times unequivocally benefits online retailers and customers, highlighting the need for a balanced management approach to order fulfillment that considers both benefits in terms of customer acquisition and downsides in terms of return costs. (c) 2024 The Authors. Published by Elsevier Inc. on behalf of New York University. This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ )

The proliferation of deep learning techniques led to a wide range of advanced analytics applications in important business areas such as predictive maintenance or product recommendation. However, as the effectiveness of advanced analytics naturally depends on the availability of sufficient data, an organization's ability to exploit the benefits might be restricted by limited data or likewise data access. These challenges could force organizations to spend substantial amounts of money on data, accept constrained analytics capacities, or even turn into a showstopper for analytics projects. Against this backdrop, recent advances in deep learning to generate synthetic data may help to overcome these barriers. Despite its great potential, however, synthetic data are rarely employed. Therefore, we present a taxonomy highlighting the various facets of deploying synthetic data for advanced analytics systems. Furthermore, we identify typical application scenarios for synthetic data to assess the current state of adoption and thereby unveil missed opportunities to pave the way for further research.

Given the ongoing "arms race" in cybersecurity, the shortage of skilled professionals in this field is one of the strongest in computer science. The currently unmet staffing demand in cybersecurity is estimated at over 3 million jobs worldwide. Furthermore, the qualifications of the existing workforce are largely believed to be insufficient. We attempt to gain deeper insights into the nature of the current skill gap in cybersecurity. To this end, we correlate data from job ads and academic curricula using two kinds of skill characterizations: manual definitions from established skill frameworks as well as "skill topics" automatically derived by text mining tools. Our analysis shows a strong agreement between these two analysis techniques and reveals a substantial undersupply in several crucial skill categories, e.g., software and application security, security management, requirements engineering, compliance and certification. Based on the results of our analysis, we provide recommendations for future curricula development in cybersecurity so as to decrease the identified skill gaps.

Deepfakes endanger business and society.Regarding fraudulent texts created with deep learning techniques, this may become particularly evident for online reviews.Here, customers naturally rely on truthful information about a product or service to adequately evaluate its worthiness.However, in light of the proliferation of deepfakes, customers may increasingly harbour distrust and thereby affect a retailer's business.To counteract this, we propose a novel IT artifact capable of detecting textual deepfakes to then explain their peculiarities by using explainable artificial intelligence.Finally, we demonstrate the utility of such explanations for the case of online reviews in e-commerce.

The increasing adoption of omnichannel strategies in recent years has led retail companies worldwide to fundamentally rethink the future role of their network of brick-and-mortar stores. One strategic option being pursued by many retailers is the transformation of the stationary store into a "smart store," augmented by various digital services. However, an essential prerequisite for the success of smart store services is high quality of the underlying data generated through the use of technologies for tracking products and customer behavior. As a means of investigating the use of machine learning to improve data quality, the present study considers the example of Radio Frequency Identification (RFID) as a technological infrastructure for tracking products in fashion retail. We examine electronic article surveillance and automated checkouts as practical use cases enabled by a classification model for the detection of product movements on the store floor. In order to identify an economically optimal configuration of the classifier, we develop a complementary service operations model that allows for determining the respective cost impact. In addition to the specific results for the considered use cases, the study thus points to a general and novel prescriptive analytics approach. (c) 2020 Elsevier B.V. All rights reserved.

Deep learning increasingly receives attention due to its ability to efficiently solve various complex prediction tasks in organizations. It is therefore not surprising that more and more business processes are supported by deep learning. With the proliferation of edge intelligence, this trend will continue and, in parallel, new forms of internal and external cooperation are provided through federated learning. Hence, companies must deal with the potentials and pitfalls of these technologies and decide whether to deploy them or not and how. However, there currently is no domain-spanning decision framework to guide the efficient adoption of these technologies. To this end, the present paper sheds light on this research gap and proposes a research agenda to foster the potentials of value cocreation within federated AI ecosystems.

Today’s working environments are subject to dynamic changes due to the proliferation of digital technologies and systems. This phenomenon poses a challenge for universities and other institutions of higher education, which are expected to adapt their course offerings to the rapidly changing demands in the labor market. The complexity of the task and the importance of speed pose an opportunity for automated data-driven methods with which the contents of study curricula can be compared, assessed, and, if necessary, adapted to the world of work. Owing to the lack of established solutions, this study presents a procedural methodology to create artifacts for analyzing, evaluating, and comparing curricula as well as job postings using topic modeling. In addition, we demonstrate the practical applicability of the methodology by the example of the IS discipline and present empirical results from the analysis of IS-related study programs in Germany.

The study considers the application of text mining techniques to the analysis of curricula for study programs offered by institutions of higher education. It presents a novel procedure for efficient and scalable quantitative content analysis of module handbooks using topic modeling. The proposed approach allows for collecting, analyzing, evaluating, and comparing curricula from arbitrary academic disciplines as a partially automated, scalable alternative to qualitative content analysis, which is traditionally conducted manually. The procedure is illustrated by the example of IS study programs in Germany, based on a data set of more than 90 programs and 3700 distinct modules. The contributions made by the study address the needs of several different stakeholders and provide insights into the differences and similarities among the study programs examined. For example, the results may aid academic management in updating the IS curricula and can be incorporated into the curricular design process. With regard to employers, the results provide insights into the fulfillment of their employee skill expectations by various universities and degrees. Prospective students can incorporate the results into their decision concerning where and what to study, while university sponsors can utilize the results in their grant processes.

Experimenting with the creative process.

FinTech (financial technology) applications are evolving at a rapid speed and are increasingly based on blockchain technology. Startups developing and offering such blockchain-based FinTech applications frequently raise capital through initial coin offerings (ICOs). Against this backdrop, it remains unclear to what extent these startups exploit the disruptive potential of blockchain. Therefore, the present article examines business models of current FinTech startups funding themselves through ICOs. The authors present their results in a consolidated business model canvas, which illustrates the current state of blockchain-based FinTech startups.

Obgleich mehrere Forscher die Wichtigkeit von Predictive Analytics für den theoretischen Erkenntnisgewinn hervorgehoben haben, werden ISTheorien der Verhaltensforschung in der Regel nicht auf ihre Vorhersagekraft getestet. Vor diesem Hintergrund soll in dem diesem Artikel zugrundeliegenden Forschungsprojekt am Beispiel der Technologieakzeptanzforschung aufgezeigt werden, dass der Einsatz von Predictive Analytics bei der Aufstellung von Theorien nicht nur wichtig ist, sondern sogar notwendig sein kann um die potenziellen Schwächen rein erklärender Modelle (insbesondere Vollständigkeit, Generalisierbarkeit und Praxisrelevanz) zu adressieren. Hierzu wird in diesem Artikel eine dreistufige Forschungsagenda zur Untersuchung und gegebenenfalls Modifizierung bestehender Modelle mithilfe von Predictive Analytics aufgezeigt.