67471 research outputs found
Sort by
Beyond Training: How Workers Discover Value in Enterprise AI
Get PDFWhile organizations continue to invest in enterprise AI, little is known about how individual employees find valuable use cases once these tools are deployed. We present an exploratory interview study of 10 experienced U.S. professionals using M365 Copilot and interpret accounts through Rogers’ Diffusion of Innovations to examine where value appears and how use cases are found and shared. Findings reveal a strong preference for informal learning methods over structured training. No participants (0/10) reported formal training as their primary way of learning; most relied on trial-and-error (8/10) and on exchanging tips with colleagues (6/10). Participants most often used M365 Copilot for note-taking/summarization, information retrieval/explanation, and writing. They also reported perceived gains in efficiency but low confidence in mastering more advanced features. The paper discusses social learning strategies and outlines implementable steps for organizations to support the discovery of high-value use cases with available enterprise AI tools
Examining the Efficacy of Multi-Theoretical Social Science-informed Deep Learning Models in Predicting Mob Outcomes
Get PDFOrganized and coordinated events, whether conducted in physical spaces (e.g., fun flash mobs, parkour, or even deviant mobs), cyberspace (e.g., collective hacking, organized propaganda campaigns), or in both spaces (e.g., electronic to face-to-face (e2f) events), represent various forms of collective action aimed at improving a group's status or condition and usually to achieve a common goal. Understanding such events requires a combination of technological and sociological approaches due to the complexity of the relationships that could exist, form, and dissolve among participating individuals. In this research, we integrate our knowledge of five social science theories that can explain such events with technical skills. We use this combination to estimate theoretical factors (both event-related and individual-related) using data collected from Meetup.com. We then train four classifiers using a deep neural network to predict the mob's outcome and rank the importance of each factor in determining the mob outcome. Results suggest that using factors related to individuals and aggregated per event made the model better classify mob outcomes (success or failure) than using event-related factors. Also, combining these factors did not affect the performance. However, all models performed better than the base model, which used raw data
Inclusive Participation in Smart City Initiatives: Mapping the Scientific Discourse
Get PDFStakeholder involvement is crucial in smart city initiatives, with growing emphasis on inclusive participation of all social groups. By means of a systematic literature review, we study how inclusive participation is conceptualized in smart city initiatives. We therefore study publications on digital and analogue participation formats in smart city initiatives and assess how various stakeholder groups are engaged. The review reveals inconsistent definitions and applications of inclusion and in-clusive participation across the literature. Drawing on these findings, we propose a structured overview and a conceptual definition of inclusive participation in smart city initiatives to support clearer frameworks for future research and implementation
Bridging Science and Medicine with the ChRIS Research Integration System
Get PDFML, AI, and cloud computing revolutionize medical imaging research, yet the lag on the clinical side in technological advancement limits the direct impact of research innovation can have on patient care. We present ChRIS, a software platform and science gateway for deployment of computational research applications across various environments, including clinical settings. It provides an end-to-end solution for research and clinical workflows, starting with DICOM data retrieval from a hospital PACS to running analysis pipelines on Kubernetes or HPC. Using ChRIS, our research center rapidly develops and deploys advanced medical software following iterative development timelines
Robust Sentiment Analysis in Service Systems: Enhancing XLNet with Adversarial Training
Get PDFIn service-intensive industries, sentiment analysis plays a pivotal role in understanding customer experiences and driving data-informed decisions. However, these systems are vulnerable to adversarial text manipulations that undermine their reliability. This paper investigates the use of XLNet, a permutation-based transformer model, for sentiment analysis in service contexts and evaluates its robustness under adversarial attacks. We propose a defense mechanism based on adversarial training with embedding-level augmentation. Experiments on diverse benchmark datasets show that while XLNet performs well on clean data, adversarial training significantly enhances robustness without sacrificing accuracy. Our findings offer actionable insights for building trustworthy sentiment analysis pipelines in service systems
An LLM-based Multi-Agent-System for the Political Assessment of Chat-LLMs
Get PDFThis paper investigates the applicability of LLM-based multi-agent systems for the political bias analysis of large language models. Following a design science research methodology, an artifact in the form of an LLM multi-agent system was created and evaluated, using the AutoGen framework. The objective was to automate and visualize bias analysis using a political questionnaire of the Bundeszentrale für politische Bildung (Federal Agency for Civic Education) in Germany. This questionnaire ("Wahl-O-Mat") contains political theses and aligns personal stances on these theses with the results of the relevant political parties in Germany. Mistral-Large-2 was chosen as an initial LLM to be evaluated with that questionnaire. The developed artifact achieved the stated goal and autonomously generated a comprehensible visualization of the results. Furthermore, the result of the automated evaluation showed a lower agreement with right-wing and right-center positions versus a higher agreement with left-wing and left-center positions
Translating and Validating Loss Aversion for Cyber Security
Get PDFLoss aversion refers to the tendency for decision-makers to weigh losses more heavily than equivalent gains. While it has been studied extensively in many domains and has a strong basis in the experimental literature, it has not been studied specifically in cybersecurity. This paper lays the groundwork for translating and measuring loss aversion in decisions relevant to cyber attackers. We conducted two experiments; the first adapts established tasks into a survey with a cybersecurity framing to assess construct validity and sensitivity to potential moderators. The second considers decisions embedded in richer cyber attack scenarios to examine loss aversions across decision stages and goal framings. Our results show that the proportion of risky action selections reliably reflects aggregate-level loss aversion, and is robust across varying payoff and probability magnitudes (including near zero and one). The stake effect significantly influences loss aversion, with greater impact at higher magnitudes. Stage effects and goal framing further modulate attacker behavior, providing insights for psychologically informed cyber defense strategies
Social Media Use as a Contributor to Marginalization in India: A Five-Year Field Study in the US and India
Get PDFThe COVID-19 pandemic accelerated a global shift to technology driven remote work, intensifying the blurring of work-life boundaries. This study investigates the differential impact on employee well-being in developed versus emerging economies. Using a five-year longitudinal study with data collected pre-COVID-19 (2019), during-COVID-19 (2020), and three years of post-COVID-19 (2021, 2022, 2023) in a single firm’s US and India offices, we examine job outcomes through the lens of the job characteristics model. Our findings reveal a stark divergence in post-pandemic recovery. While US employees’ job strain and job satisfaction returned to pre-pandemic levels, their Indian counterparts experienced sustained job strain and reduced job satisfaction recovery. These persistent negative outcomes demonstrate that the impact of using technologies in India has led to them being more marginalized driven by the new realities of technology-dependent work. That Indian workers experience significantly higher job strain, which has not returned to pre-COVID levels, represents a grave concern. This serves as an urgent call for scientific and managerial attention to the problem of emergent marginalization due to increased social media and other technology use