67471 research outputs found
Sort by
Counting Species of Ideas: A Bayesian Capture–Recapture Ecology Framework for Estimating LLM Novelty
Get PDFLarge Language Models (LLMs) are increasingly used for ideation tasks across domains, ranging from product development to marketing and creative writing. Yet, we lack principled methods to quantify their genuine capacity for novelty and ideation. LLMs derive their generative potential from extensive training data, model architectures, and optimization objectives. These factors collectively define a large yet bounded ideation space. However, standard evaluation methods—typically centered around output-level novelty or diversity—only capture a limited view of this broader ideation landscape. To overcome this limitation, we propose shifting the evaluation focus from isolated outputs to this underlying ideation space. Effectively exploring this space involves answering foundational questions: How expansive is this space? How many unique ideas can the model potentially generate? By adapting capture-recapture (CR) theory from ecology, we introduce an estimation framework tailored to the generative behavior of LLMs and infer the unseen idea space beyond observed samples. Validation through asymptotic extrapolation confirms the reliability of our framework, which offers a principled approach to understanding and comparing the innovation capacities of LLMs
A Multi-Task Learning Approach for Predicting Capacity Expansion Timing and Requirements in Colocation Datacenters
Get PDFColocation data centers are essential infrastructure for the digital economy, supporting scalable and secure operations across industries. With the global colocation market exceeding $218 billion—fueled by data growth, technology, and economic expansion—intelligent long-term capacity planning is critical. Traditional reactive methods often miss localized demand shifts, leading to under-provisioning or costly over-investment. A key challenge lies in forecasting when and how much capacity will be needed across infrastructure segments, such as power delivery domains. We propose a multi-task learning framework to jointly predict the timing and magnitude of future capacity expansions. Our hybrid Transformer-based architecture integrates static and temporal features, such as facility telemetry, sector metadata, macroeconomic indicators, and sentiment signals, into a unified temporal embedding space with a static feature layer. It generates dual outputs: a binary classifier for expansion events and a conditional regressor for size. By modeling long-range dependencies and uncertainty, our approach enables accurate, adaptive forecasts that support proactive procurement, smarter resource allocation, and improved infrastructure agilit
Towards a Blueprint for Practitioners to Enhance Digital Awareness: Preliminary Observations on AI-Related Workplace Stressors
Get PDFThe growing integration of Artificial Intelligence (AI) in the workplace introduces both positive and negative stressors that affect employee well-being and organizational efficiency, requiring effective management of these dynamics. Empirical evidence on the effectiveness of digital awareness training in mitigating AI-related stress remains scarce. This pilot study develops a digital awareness training and evaluates its impact on AI-related stress appraisal, including techno-distress, techno-eustress, and attitudes towards AI (fear vs. acceptance). Results from a heterogeneous sample of employees (n = 32) from small and medium-sized enterprises indicate a significant reduction in techno-overload, with tendencies towards improved work facilitation and information literacy. These results offer an initial, evidence-based blueprint for digital awareness interventions, offering actionable insights for reducing techno-distress and promoting digital literacy. Further longitudinal and repeated-intervention studies are needed to validate and sustain these effects over time
Designing Effective Empathy in AI Agents: An Empirical Study of User-Centric vs. Situation-Centric Approaches between Human and AI Agents
Get PDFAs generative AI systems play an increasing role in emotional support, scholars have raised concerns about discomfort, inauthenticity, and expectancy violations resulting from AI's empathic responses. Drawing on verbal person-centeredness theory, we propose User Centric Empathy (UCE: emotion-focused and validating) and Situation Centric Empathy (SCE: context-focused and redirecting) to identify a more effective AI empathy approach. Across two experiments, we investigate how empathy type (UCE vs. SCE) and agent type (human vs. AI) interact to shape user experience. The results indicate that, when expressed by an AI agent, situation-centric empathy (SCE) emerges as a more appropriate empathy strategy, as it reduces discomfort and inauthenticity. Interestingly, when blame is attributed to one’s own error rather than to external sources, the type of empathy expressed by the AI agent exerts no significant effect. These results highlight that the effectiveness of AI-delivered empathy depends less on mimicking human-like responses and more on adopting an appropriate empathy approach, showing that superficial mimicry cannot foster authentic relational outcomes
Hiring Tomorrow’s Talents: How Generative Artificial Intelligence Transforms Human Resources Recruitment
Get PDFThe global talent shortage has become a universal challenge, prompting practitioners and researchers to explore digital innovations as potential solutions for acquiring the right talents. However, the role of emerging technologies like generative artificial intelligence (AI) in human resources (HR) remains largely uncharted territory. This article investigates generative AI’s transformative potential to augment recruiters’ daily operations. Through a qualitative interview study, we derive and illuminate the opportunities of generative AI within the recruitment domain, shedding light on its promising opportunities but also addressing inherent challenges. The findings of this study propose a theoretical model of generative AI in recruitment and how it empowers recruiters in their daily tasks to recruit tomorrow’s talents
Uneven but Better? Unequal AI Access Leads to Greater Dominant Style Differences and Enhanced Team Communication Effectiveness
Get PDFThe widespread use of generative AI (GenAI) is seen as beneficial for team collaboration, yet full access for all members is often impractical in real-world settings. This study investigates how varying levels of GenAI integration influence team communication dynamics: no access (team members do not use AI), unequal access (only some members use AI), and full access (all members use AI). In a laboratory experiment with 60 two-person teams, all teams first performed a task without AI, then were randomly assigned to either the unequal or full access condition. Under unequal access, AI users adopted more dominant communication styles, creating greater style differences that, in turn, enhanced team communication effectiveness compared to both no access and full access. This research provides theoretical insights into the effects of varied GenAI integration structures on human-AI collaboration and offers practical guidance for optimizing team design in human-agent systems
Engineering Better Requirements: Understanding the Impact of GenAI on Task Performance and Quality in Requirements Engineering
Get PDFIn order to deliver high-quality software systems, organizations utilize Requirements Engineering (RE) as a foundation for aligning development with stakeholder needs. With advances in Generative Artificial Intelligence (GenAI), potential emerges to augment RE processes to improve both task completion time and requirements quality. As GenAI-powered tools become more common in industry practice, our research examines under which circumstances GenAI supports RE tasks. Through our online experiment with 41 RE professionals from a manufacturing company, we demonstrate that GenAI assistance significantly reduces task completion time across different complexity levels and improves quality, particularly in simpler tasks. These results indicate that while GenAI effectively enhances RE efficiency in all contexts, its contribution to quality varies with task complexity. This suggests that organizations should strategically implement GenAI tools in RE workflows, recognizing both their productivity benefits and the continued importance of human expertise for more complex requirement scenarios
Understanding Computer Science Students' Career Fair Experiences: Goals, Preparation, and Outcomes
Get PDFThe technology industry offers exciting and diverse career opportunities, ranging from traditional software development to emerging fields such as artificial intelligence, cybersecurity, and data science. Career fairs play a crucial role in helping Computer Science (CS) students understand the various career pathways available to them in the industry. However, limited research exists on how CS students experience and benefit from these events. Through a survey of 86 students, we investigate their motivations for attending, preparation strategies, and learning outcomes, including exposure to new career paths and technologies. We envision our findings providing valuable insights for career services professionals, educators, and industry leaders in improving the career development processes of CS students
Designing Conversational Agents to Support Learning from Scientific Graphs
Get PDFThe need for accountability in research and an informed policy-making and society has increased demand for publicly accessible research data. However, the tendency in the scientific context to rely on graphs to illustrate findings can be challenging for users lacking domain expertise. Rooted in Conversation theory, this study explores using conversational agents to help users interpret such graphs. Using design science research, we develop and test a prototype conversational agent, addressing previously identified challenges like accessibility as well as language and education barriers in graph interpretation. While users with conversational agent access in a large-scale experiment do not demonstrate improvements in objective learning success, they report higher perceived learning, perceived empowerment and user experience, with reduced cognitive load. This research highlights the potential of large language model-based agents to improve access to complex data, offering insights into science communication as well as educational agent design and their limitations
Designing Educational Games for Financial Literacy and Bias Awareness: Reflections on Challenges and Opportunities
Get PDFFinance is an important yet underexplored context in the development of prescriptive knowledge for education interventions addressing cognitive biases. Previous research indicates that games are a promising tool for cognitive bias mitigation, but our understanding of how to effectively design games to enhance bias awareness remains limited, particularly in financial education. This study explores challenges and opportunities therein based on seven focus groups conducted with upper secondary school students after playing a financial education game. A thematic analysis of the data indicates that students consider finance uncertain and complex, and wish for games to match this reality through simulations that are challenging and applicable to real-life contexts. Based on these, we propose four design principles to guide the design of effective financial education information systems and games—adopting a constructivist approach to bias education; supporting agency through scaffolding; contextualizing decision-making; and paying special attention to learners' attention and motivation limitations