67471 research outputs found
Sort by
Navigating the Generative AI Adoption Dilemma: Pathways and Trade-offs through a Data Ecosystem Perspective
Get PDFThe rapid rise of Generative AI (GenAI) chatbots such as ChatGPT and Claude presents organizations with a strategic dilemma: adopt powerful external solutions or invest in developing secure, internal alternatives. This study conceptualizes the "GenAI adoption dilemma" and examines it through the lens of internal data ecosystems. Based on qualitative case study research, including 46 interviews with senior professionals in a multinational consultancy firm, we identify three strategic pathways: internal, external, and hybrid chatbot adoption. We analyze them across five criteria: cost, risk, time to market, data ecosystem integration, and competitiveness. We develop a framework to guide decision-makers in navigating this emerging challenge. Our findings primarily contribute to GenAI chatbot adoption literature by framing it as a systemic organizational decision rather than a purely technological, they extend data ecosystem theory by empirically linking GenAI chatbots adoption to internal infrastructure, governance, stakeholders, and data practices, and provide insights for IT adoption and outsourcing literature
Reinforcement Learning for Adversarial Systems Using Relational Observations
Get PDFThis paper investigates the integration of relational observations with the reinforcement learning (RL) framework for improved generalization capability. A hide-and-seek simulation environment is designed in Unity for proof-of-concept demonstration. Two observation representations—relational (analogical) and standard positional—are designed to evaluate agent learning and generalization capabilities. Agents are trained using the Proximal Policy Optimization (PPO) algorithm in a random-room environment and tested in both the random-room environment and a novel environment with greater spatial complexity and path obstructions. Comparative studies indicate that relational representation of objects in the adversarial environment could potentially improve the generalization capability of RL agents to novel and complex environments. Cross-testing results also suggest that relational observations may enhance agents’ effectiveness in pursuit and evasion tasks in adversarial environments
Solving Three-phase AC Infeasibility Analysis to Near-zero Optimality Gap
No full textRecent works have shown the use of equivalent circuit-based infeasibility analysis to identify weak locations in distribution power grids. For three-phase power flow problems, when the power flow solver diverges, three-phase infeasibility analysis (TPIA) can converge and identify weak locations. The original TPIA problem is non-convex, and local minima and saddle points are possible. This can result in grid upgrades that are sub-optimal.To address this issue, we reformulate the original non-convex nonlinear program (NLP) as an exact non-convex bilinear program (BLP). Subsequently, we apply the spatial branch-and-bound (sBnB) algorithm to compute a solution with near-zero optimality gap. To improve sBnB performance, we introduce a bound tightening algorithm with variable filtering and decomposition, which tightens bounds on bilinear variables. We demonstrate that SBT significantly improves the efficiency and accuracy of Gurobi's sBnB algorithm. Our results show that the proposed method can solve large-scale three-phase infeasibility analysis problems with >5k nodes, achieving an optimality gap of less than 10e-4. Furthermore, we demonstrate that by utilizing the developed presolve routine for bounding, we can reduce the runtime of sBnB by up to 97%
Introduction to the Minitrack on Design, Implementation, and Management of Digital Government Policies and Strategies
No full textGenerational Gender Differences in the Role of Time in 360-Virtual Shopping
Get PDFGenerational differences shape how individuals engage with digital technologies. Immersive shopping environments, such as 360-virtual stores, impair consumer attention, often leading to a loss of temporal awareness and predisposing to unplanned purchases. Given that individuals’ sense of time passing varies with age and gender, we set up a study in a 360-degree virtual store to study how time distortion and telepresence in a 360-virtual environment influence impulse buying behavior differently across generational cohorts and generational gender groups. The results show that time distortion significantly increases impulse buying, but our cross-generational analysis suggests that this effect is only significant for Millennials. When we split generational cohorts across gender, we find that the effect of time distortion on impulse buying is significant also for Generation Z and Generation X, but for females only. The effect is non-significant for Baby Boomers, while telepresence further boosts the effect for Millennial women
From Automation to Collaboration: A Systematic Review of AI Use in Assessment Across Critical Infrastructure Sectors
Get PDFAssessments are used to help gather and analyze information to inform processes and outcomes and are rapidly being reshaped by AI. This systematic review investigates where, why, and when AI is used across the assessment life-cycle and further considers its core functions, design elements, and the ways users engage with them Thirty-eight peer-reviewed studies met our inclusion criteria, each embedding artificial intelligence directly into the assessment process. Together, government facilities and healthcare settings accounted for more than 70% of all documented use cases. Across sectors, the prevailing role of AI was that of a digital assistant, streamlining knowledge capture and evaluation supporting assessment in its role as an expert with a focus on goal-oriented collaboration. These patterns illuminate both the breadth of adoption and the potential of AI as an augmentative partner, offering a roadmap for future assessment design and research
Bridging NASSS Domains and Resilience Pillars to Mitigate LLM Risk in Digital Government
Get PDFPublic sector workflows are rapidly moving to deploy Large Language Models (LLMs). However, today’s governance standards stop at high-level duties without concrete escalation logic. This study introduces NASSS-RE-LLM, a socio-technical governance schema that maps LLM risks to NASSS domains, while pairing each with a pillar of Resilience Engineering. Using a scenario-based methodology, we test mitigation strategies in simulated public hospital settings. Results from expert panels show that the framework effectively surfaces risks, and experts perceive it as clear and adaptable. Practical implementation may be limited by resource constraints and future research is needed to test the framework in real-world pilots. The findings suggest that embedding governance logic into public sector workflows can enhance LLM adoption, with potential broader applicability across digital government contexts
Neurodivergent Citizens in Search of Wellbeing: Addressing Barriers through Digital Means
Get PDFThis paper investigates how digital technologies can support neurodivergent individuals in navigating urban environments and participating in democratic processes, drawing on in-depth interviews with neurodivergent residents of Warsaw, Poland. Participants described a range of cognitive, sensory, and communicative challenges when engaging with urban infrastructures – from navigating transit systems to accessing online public services. Many public spaces or services were found to be inaccessible or overwhelming, contributing to exclusion from civic life. At the same time, interviewees identified the transformative potential of well-designed digital interventions, such as personalized navigation aids, simplified user interfaces, and platforms that allow for asynchronous civic engagement. Framed through the lens of cognitive justice, this study highlights the need for participatory, neurodiversity-affirming approaches in the design of digital government and smart city initiatives. By centering neurodivergent perspectives, the paper contributes to the discourse on inclusive urbanism and proposes strategies that uphold the right to the city for cognitively diverse citizens
Rethinking Knowledge Management in the Era of LLMs: A Knowledge Management Model for Mitigating LLM Hallucination
Get PDFLarge Language Models have the potential to greatly enhance organizational knowledge management practice. However, along with this promise comes the risk of hallucination. LLM hallucination can proliferate rapidly, threatening the quality of knowledge available to employees. This paper proposes a conceptual model to guide the mitigation of hallucination-related knowledge risks in LLM-powered knowledge management systems. Building on the Knowledge Management Cycle Model, the updated framework integrates practical hallucination mitigation strategies, including Prompt Engineering, Retrieval-Augmented Generation, Automated Verification, and Human Feedback, into each stage of the cycle. While the model remains general and conceptual, it provides a foundational roadmap for organizations to adapt knowledge practices in the age of LLMs and offers clear directions for future empirical research and refinement