67471 research outputs found
Sort by
Reactions for Geopolitical Shocks on Online Labour Platform: Empirical Evidence Under Ukraine-Russia War
Full text linkAs digital labor platforms play an increasingly central role in the global workforce, their governance decisions—especially in response to geopolitical crises—have profound implications for market outcomes and worker livelihoods. Despite their growing importance, little is known about how platforms strategically respond to external shocks and how these decisions reshape opportunity structures for different types of freelancers. This study exploits a natural experiment arising from a platform governance decision to remove all Russian freelancers in May 2022, following the onset of the Russia–Ukraine war. Drawing on a granular panel dataset of 428,401 daily work records linked to 80,141 freelancer profiles, we apply a regression discontinuity in time (RDiT) design to estimate the causal effects of this exogenous labor supply shock. We find that the intervention significantly increased average earnings for the remaining workforce, primarily through higher income per task rather than an increase in task volume. By situating the analysis within a platform governance perspective, this study contributes to research on labor market shocks, platform strategy, and algorithmic reallocation. Our findings show that platform responses to geopolitical risk can reinforce structural inequalities through selective opportunity redistribution, highlighting the strategic role of platform governance in shaping post-shock labor market dynamics
Does Private Equity Hurt or Improve Healthcare Value? New Evidence and Mechanisms
Full text linkWhat is the impact of private equity (PE) investment on healthcare value? Does such investment hurt care value, and if so, can this effect be mitigated through the adoption of health IT systems? These are important questions given the growing presence of PE firms in the healthcare sector. Stakeholders, including policy makers, care providers, and patients, need to understand the likely impact and whether PE ownership aligns well with their interests. This study examines the impact of PE investment on healthcare value, focusing on the complex interplay between cost reduction and patient care quality. Using hospital-level data in the U.S. from 2008-2020, we estimate changes in healthcare value, defined as the balance between clinical resource use and patient outcomes, following PE acquisition. We find that overall healthcare value declines after PE investment. However, our empirical evidence also reveals that IT-enabled health information sharing plays a mitigating role. Hospitals with strong information-sharing capabilities experience both greater cost efficiencies and improvements in care quality, ultimately leading to higher healthcare value after PE investment. Moreover, we find that the type of information sharing matters: improvements in care quality are primarily driven by hospital-to-ambulatory provider information sharing
AI-Enabled Forensic Risk Assessment: TRAP-18 System Architecture and Proof of Concept
Full text linkThis study explores the development and validation of an AI-enabled forensic risk assessment system using the Terrorist Radicalization Assessment Protocol-18 (TRAP-18) as a proof of concept. Building on prior work demonstrating large language models’ (LLMs) ability to code TRAP-18 indicators with expert-level reliability, this project integrates structured professional judgment (SPJ) methodology with advanced AI frameworks, including LangChain and LangGraph. A prototype system was constructed to simulate a full TRAP-18 evaluation, encompassing indicator coding, justification generation, Bayesian probability estimation, and narrative risk formulation. Results demonstrate high consistency with human raters on proximal warning behaviors and strong agreement across multi-model workflows. The architecture emphasizes transparency, reproducibility, and bias mitigation, highlighting AI’s potential to augment forensic practice through structured reasoning, hypothesis testing, and scalable data integration. Beyond TRAP-18, this framework offers a pathway toward AI-assisted applications in general violence risk assessment, reinforcing human-AI collaboration in forensic psychology
Small Language Models for Curriculum-based Guidance
Full text linkThe adoption of generative AI and large language models (LLMs) in education is still emerging. In this study, we explore the development and evaluation of AI teaching assistants that provide curriculum-based guidance using a retrieval-augmented generation (RAG) pipeline applied to selected open-source small language models (SLMs). We benchmarked eight SLMs, including LLaMA 3.1, IBM Granite 3.3, and Gemma 3 (7–17B parameters), against GPT-4o. Our findings show that with proper prompting and targeted retrieval, SLMs can match LLMs in delivering accurate, pedagogically aligned responses. Importantly, SLMs offer significant sustainability benefits due to their lower computational and energy requirements, enabling real-time use on consumer-grade hardware without depending on cloud infrastructure. This makes them not only cost-effective and privacy-preserving but also environmentally responsible, positioning them as viable AI teaching assistants for educational institutions aiming to scale personalized learning in a sustainable and energy-efficient manner
Using Google’s Natural Language Model to Measure Growth of Knowledge in Information Systems Research
Full text linkThe goal of this paper is to propose a new artificial-intelligence (AI) driven method to evaluate how well the information systems (IS) field engages with other disciplines in the process of building IS knowledge. The proposed method combines the veracity and objectivity of quantitative scientometric methods with the semantic depth and interpretive validity of qualitative content analysis methods, both building on theories of citations and disciplinarity. The results find that the IS field relies mostly on reviewed and perfunctory citation functions that do not truly engage with previous research. This evaluation presents a wake-up call to the field to better leverage and engage with theories from previous research. It also showcases the scientometric bases for enhancing the originality of IS research and help the field become intellectually and socially influential. Keywords: Disciplinary theory, citation theory, artificial intelligence (AI) and natural language processing, information systems (IS) knowledge
Introduction to the Minitrack on Human-in-the-Loop Hybrid Augmented Intelligence Systems
No full textIs Visual Complexity Always Better? Evidence from Cover Images in Online Crowdfunding
No full textIn the highly competitive environment of crowdfunding platforms, cover images have become a crucial force in capturing the attention of potential backers. However, effective visual communication strategies for cover images remain underexplored. To address this gap, we investigate the dual impact of visual complexity in cover images. We divide visual complexity into two dimensions: feature complexity at the pixel level and design complexity at the object level. Based on the Elaboration Likelihood Model (ELM), we propose that these dimensions of visual complexity have differentiated effects in highly competitive crowdfunding contexts: feature complexity helps enhance project performance, whereas design complexity exerts a negative influence. Empirical analysis of large-scale project data from over 1.2 million loan-based crowdfunding campaigns validates these arguments. The study enriches our understanding of how visual factors shape crowdfunding outcomes and offers practical guidance for fundraisers to optimize cover images design
What We Talk About When We Talk About DAOs: An Integrated Socio-Technical Framework
No full textDecentralized Autonomous Organizations (DAOs) integrate blockchain-based automation with novel forms of collective governance, yet research on them remains fragmented across technical and organizational silos, hindering a comprehensive understanding of DAOs as socio-technical systems. To bridge these gaps, we first conduct a systematic umbrella review of 12 prior surveys to map research themes and persistent gaps. Based on this analysis, we propose a novel, three-layer framework that explicitly links (i) technical artefacts (the infrastructure, e.g., tokens, smart contracts), (ii) governance logics (the rules, e.g., incentives, consensus mechanisms), and (iii) organizational manifestations (the outcomes, e.g., proposals, votes). By making cross-layer dependencies explicit, the framework enables more holistic theorizing, supports comparative empirical work, and provides a diagnostic tool for practitioners dealing with design trade-offs between decentralization, efficiency, and participation