AIS Electronic Library (AISeL)
Not a member yet
72426 research outputs found
Sort by
Teaching Data Preparation to Non-technical Audiences Using Tableau Prep Builder
Get PDFAn increasing number of Information Systems courses aim to improve students\u27 data analytics skills. Teaching such skills requires teaching data preparation practices. These skills were traditionally taught using tools targeted at technically trained audiences with coding abilities. However, there is an increasing demand to learn data preparation skills from non-technical audiences in business school programs. To this end, Tableau-Prep Builder (TPB) was used to teach Extract Transform Load at an introduction to analytics master’s level course in a business school program at the Université du Québec à Montréal in winter 2025. We note that TPB belongs to the same publisher as Tableau Desktop, the popular data visualization tool
Understanding Vaccine Awareness and Digital Engagement among African American Parents: A Survey-Based Approach
Get PDFAfrican American children continue to experience lower childhood vaccination rates (Immunizations and Black/African Americans | Office of Minority Health, 2025), driven by factors such as vaccine hesitancy, historical mistrust, and misinformation. Despite the availability of vaccines, rates of vaccination remain below recommended levels, particularly in underserved communities (Hill, 2024). This research-in-progress aims to explore the underlying awareness gaps and digital behaviors that may influence how African American parents, and future parents, make vaccine related decisions for their children. Research has shown that digital health interventions are effective tools in increasing health knowledge and engagement, especially amongst ethnic and cultural minorities (Radu et al., 2023). We employ a survey-based methodology to assess baseline knowledge of vaccine schedules, perceptions of vaccine safety, and openness to digital tools for health education. The survey also captures digital literacy, mobile app usage patterns, and attitudes toward gamification in health contexts. Early findings are expected to provide a detailed understanding of informational needs, behavioral barriers, and the potential acceptance and utilization of technological solutions. This study is part of a broader research agenda to design inclusive and culturally relevant digital health interventions. By grounding our future app development in data from impacted communities, we aim to improve the effectiveness and trustworthiness of public health messaging. Improved public health messaging is especially important for populations that exhibit distrust of health institutions due to historical harm. Our preliminary results will provide insights into how African American parents and future parents perceive vaccination content and strategies to improve digital engagement in childhood healthcare. Preliminary results will also help inform design choices for future mobile health tools aimed at broader populations
Beyond the Crowd: A Literature Review of AI’s Impact on Crowdsourcing Systems
Get PDFCrowdsourcing harnesses the collective intelligence of diverse individuals across organizational boundaries. Traditionally, job posters define the task requirements, post jobs on crowdsourcing platforms, and assess the quality of the work. Recent advancements in artificial intelligence have begun to streamline the entire process. Artificial intelligence can assist job posters in clearly defining job requirements, distributing tasks to qualified workers by matching job requirements with their backgrounds, and evaluating the quality of contributions. The advantage of AI in crowdsourcing is its compatibility with various crowdsourcing models (e.g., idea generation) (Dissanayake et al., 2025). In the idea generation model, AI can group and highlight similar generated ideas as submissions roll in and evaluate their quality based on novelty and feasibility in achieving the goals. Similarly, AI can detect abnormalities (e.g., speedy response times) in the microtasking model and calculate the task error rate. Scholars in information systems (IS) and other disciplines have investigated the role of AI in various fields, demonstrating its potential in handling complex tasks. While current research has explored the use of AI in crowdsourcing, the findings are scattered and vary among crowdsourcing models. The literature review of AI in crowdsourcing provides a clear understanding of how AI transforms crowdsourcing operations and identifies research gaps. Therefore, we organize and analyze the current use of AI in crowdsourcing operations and its impact on these operations. Our work employs the Input-Process-Output (IPO) model, a framework widely used in management research, to examine the current application of AI in crowdsourcing. This framework allows us to decompose the crowdsourcing process into three subprocesses: input, process, and output (Ghezzi et al., 2017). The “input” stage involves defining tasks that workers should perform. Building on this foundation, the “process” stage focuses on how job posters manage the crowdsourcing session (e.g., organizing the submissions during the session), which is necessary for the “output” stage, where solutions are evaluated and selected. In a crowdsourcing process, AI plays a unique role, and how and where AI is used in the process determines the outcomes for each stage (e.g., enhanced clarity of job descriptions). Understanding the outcomes of using AI in crowdsourcing can help us assess the impact of AI on the crowdsourcing process. Therefore, we propose the following research questions: “How is AI applied across the stages of a crowdsourcing process among various crowdsourcing models? ” (RQ1) and “What are the impacts of AI usage across the stages of the crowdsourcing process among various crowdsourcing models? ” (RQ2). In conclusion, this literature review offers a clear understanding of the current application of AI in the crowdsourcing process by breaking down the process into three subprocesses, enabling an examination of the nuances of AI\u27s impact on each subprocess
Novel model for Better Segmentation
Get PDFDigital imaging techniques have advanced significantly since the 1960s (Gonzalez & Woods, 2007), with computer algorithms being employed to enhance contrast, encode intensity levels, and enable efficient object recognition. These advancements have revolutionized various fields such as X-ray interpretation, medical image analysis, and satellite imaging. Image segmentation, as a critical preprocessing step, is essential for tasks ranging from precise disease diagnosis (e.g., tumor localization in CT scans) to environmental monitoring (e.g., land cover classification in satellite imagery). The state-of-the-art models for image segmentation often leverage information from multiple scales, with the U-Net architecture being one of the most prominent examples (Szegedy et al., 2015). U-Net’s distinctive U-shaped architecture utilizes skip connections to merge high-level semantic feature maps from the decoder with corresponding low-level detailed feature maps from the encoder (Smith & Doe, 2022). Combined with powerful data augmentation techniques, U-Net maximizes the use of limited annotated samples. However, traditional segmentation methods often struggle with complex feature extraction and computational efficiency, especially in scenarios with limited annotated data or resource-constrained environments. To address these challenges, attention mechanisms have emerged as a powerful tool to enhance model sensitivity to task-relevant features. Among them, the Efficient Channel Attention (ECA) mechanism stands out due to its ability to adaptively recalibrate channel-wise feature responses without dimensionality reduction, significantly reducing computational overhead while maintaining performance Prior studies have demonstrated the effectiveness of ECA-integrated architectures in medical imaging. For instance, ECAU-Net improved fetal ultrasound cerebellum segmentation (Brahmankar et al., 2022) and enhanced performance in coronary artery segmentation and three-dimensional reconstruction (Brahmankar et al., 2022). Yet, these applications remain confined to the medical domain, with limited exploration in non-medical contexts. Building upon the success of U-Net in brain tumor image segmentation (Doe & Smith, 2024) and cerebellum segmentation for clinical diagnosis (Murugan & Karuppiah, 2022), we are among the first study to introduces the first extension of ECA-enhanced U-Net architecture to general image segmentation tasks beyond healthcare. We term this approach ECAU-Net, which integrates the Efficient Channel Attention (ECA) mechanism into U-Net’s skip connections to dynamically prioritizes informative channels across scales, enabling robust segmentation in diverse scenarios such as industrial defect inspection and agricultural crop monitoring, while preserving computational efficiency. By applying the encoder-decoder architecture, ECAU-Net efficiently locates segmentation results, making it a powerful backbone for various segmentation applications. As a result, the improved U-Net demonstrates a significant enhancement in segmentation accuracy. During the experimental evaluation, the improved model was trained and systematically assessed, with results showing a consistent 2% improvement in key metrics—mean intersection over union (mIoU), mean pixel accuracy (mPA), precision, and recall—compared to the traditional U-Net and faster convergence in training loss. These improvements underscore the efficacy of the proposed approach, demonstrating that the addition of the ECA attention mechanism leads to more precise and reliable segmentation outcomes
AI and Higher Education: Navigating the New Frontier
Get PDFWith increasingly powerful generative artificial intelligence (AI) tools now widely available, post-secondary institutions are struggling to keep up with a rapidly evolving landscape. While much has been discussed about the potential impact of this technology, there is still limited empirical data on how students, educators, and administrators are integrating AI in teaching and learning contexts. In this presentation, we share key findings from a multi-methods study conducted by The Conference Board of Canada, on behalf of the Future Skills Centre (FSC). The study includes a national survey of postsecondary students (N=2,401) and educators (N=402), as well as interviews with individuals leading responses to AI in higher education institutions (N=42). We found that frequent usage was not widespread among students, with 20 per cent of students reporting using generative AI most or all of the time. Usage varied significantly across students in different sociodemographic groups and fields of study. Power users – those who report using generative AI most or all of the time – had similar levels of concern as non-users about the potential drawbacks of generative AI, despite having more favourable attitudes toward its use. We also found an association between frequency of use and better learning experiences and outcomes, but the mechanisms and conditions under which this occurs need to be further investigated. Among educators, we found that most have neither explicitly permitted nor prohibited student use of AI tools. Notably, 80% reported not receiving any formal guidance or training from their institutions. There is a strong demand for professional development in this area, with educators seeking training for both themselves and their students. Educators who use generative AI more frequently tend to be more optimistic about its potential, although they remain wary of its ethical implications and possible threats to the integrity and reliability of knowledge. Conversations with institutional leaders revealed a wide range of perspectives on AI, from views that it has radically transformed the role of the teacher, to skepticism about its overall impact. Many leaders expressed enthusiasm for AI as a tool for enhancing higher-order learning. These findings have important implications for various post-secondary stakeholders, particularly instructors and administrators looking to integrate AI into educational environments. We conclude with recommendations focused on fostering critical literacy, ensuring transparency and accountability in AI use, and promoting equity in access to AI tools and training
When AI Gets It Wrong: Investigating the Effects of Output Inaccuracy on User Trust and Continuance Intentions
Get PDFArtificial Intelligence (AI) systems are increasingly embedded in high-impact decision-making contexts such as healthcare, hiring, and financial services. While much of the literature has focused on explainability and fairness, fewer studies have examined how users respond when AI outputs are inaccurate. Given that real-world AI is probabilistic and often imperfect, understanding how output inaccuracy affects trust and continu-ance intentions is critical for designing robust and trustworthy systems. Drawing on trust calibration theory (Lee & See, 2004) and information systems continuance models (Bhattacherjee, 2001), this study investigates how users adjust their trust after experiencing incorrect AI out-puts and whether transparency mechanisms such as confidence scores or uncertainty indicators can mitigate trust erosion. We theorize that output inaccuracy undermines perceived system reliability, reducing both trust and intention to continue using the AI. However, if the system acknowledges its limitations or flags uncer-tain results, users may calibrate their trust more appropriately and remain engaged. We propose a 2x2 between-subjects experiment involving participants interacting with a decision-support AI system in a simulated task (i.e., evaluating résumés). The two factors are: (1) AI Output Accuracy (accurate vs. inaccurate) and (2) Transparency Mechanism (present vs. absent). Dependent variables include trust, perceived reliability, and continuance intention. The study will assess changes in trust over time and explore whether users override or defer to AI recommendations after errors occur. Controls variables will include in-dividual characteristics such as algorithm aversion and propensity to trust. This research offers three contributions. First, it extends trust-in-AI literature by examining user responses to inaccuracy. Second, it examines transparency mechanisms that may buffer the negative effects of errors on trust. Third, it provides practical guidance for AI designers seeking to build systems that retain user trust de-spite the inevitable failures in complex, real-world AI applications
Deepfake Detection Using CNN-LSTM and Multimodal Analysis: A Hybrid AI Approach
Get PDFThe quick evolution of artificial intelligence and generative models has made it possible to produce hyper-realistic fake media, also known as deepfakes. Although such technology has potential uses in entertainment and education, it is a major threat when used for misinformation, identity theft, or defamation. This research fills the critical need for strong and scalable detection mechanisms by investigating multimodal deep learning methods to identify deepfakes in real-world applications. This work provides an in-depth comparison of different deep learning models, including convolutional neural networks (CNNs), long short-term memory (LSTM) networks, and transformer models for the detection of fake video content. The uniqueness of this work lies in its focus on multimodal information, examining visual and auditory features for better detection. The models are trained and tested on benchmarking datasets like FaceForensics++ and the Deepfake Detection Challenge (DFDC), maintaining diversity and realism while testing. This work\u27s experimental results prove that multimodal methods far exceed unimodal models, especially in identifying subtle forgeries under adverse conditions like compression and occlusion. Out of the configurations tried, a hybrid model integrating ResNet-50 for visual frames and Bi-LSTM for audio streams achieved an accuracy rate of 94.6% on the DFDC test set and exhibited excellent generalizability. In addition, our research identifies key issues in actual deployment, including adversarial attacks, dataset bias, and computational cost. To address these, we introduce methods such as data augmentation, domain adaptation, and model compression without severely degrading performance
Do Perceptions of Digital Technology Advances Affect Knowledge-Hiding Behavior? A Study in the United States
Get PDFBackground: Diverging from prior information system (IS) research that investigates the impact of information technology (IT) usage and its correlates of technostress and techno-insecurity, we examine the role of employees’ perceptions of IT advancement and its relationship to their work behaviors. Drawing on conservation of resources (COR) theory, we propose that if employees perceive their jobs may be replaced by smart technology, artificial intelligence, robotics, and algorithmic (STARA) technologies—that is, if they have a high perceived STARA awareness—then their knowledge-hiding behavior will be elicited via feelings of job insecurity.
Method: We conducted a two-wave survey in the United States with an interval of one month using Connect-Cloud Research’s sampling system. Referencing the Automation Risk Score from the website Will Robots Take My Job? (https://willrobotstakemyjob.com/), we obtained a sample of 165 participants of a wide distribution of possibilities for job replacement by STARA technologies.
Results: Our study results provide external validity for the STARA awareness scale and support the proposed hypotheses. Specifically, we find that STARA awareness is positively related to feelings of job insecurity and to knowledge-hiding behavior via feelings of job insecurity.
Conclusion: This study adds to the literature on technological development and knowledge management by highlighting that employees’ perceptions of IT advancement may have a consequential negative impact on their work behavior. We suggest that organizations in most Asia Pacific economies, which are still in earlier phases of AI integration, can leverage the temporal gap to implement proactive measures to mitigate negative consequences of such perceptions. Also, given that fear about uncertain technological changes can prompt self-serving responses, organizations should prioritize transparent communication to alleviate employees’ job insecurity and lower their knowledge-hiding behavior. Management should invest in human resource initiatives to address these concerns, especially for employees who are facing imminent displacement by new technologies
Digital Business & Consumer Insights on Emerging Social Media Platforms
Get PDFSocial media, enabled by various digital technologies, e.g., data analytics, blockchain, artificial intelligence, and augmented or virtual reality, has increasingly played a central role of monitoring, responding, optimizing, and influencing consumer behavior in digital business (Chou et al., 2025; Do et al., 2025). Since 2017, the immediate popularity of Douyin in China and later TikTok in the world brought the new paradigm of social media platforms by leveling down the threshold of video production and directing platform competition from celebrity/content centric to algorithm centric. Such a paradigm shift suggests new challenges and opportunities in digital business, i.e., constructing sustainable relationships among “content producers -- content consumers/customers -- businesses/brands”. The latest development of AI technologies also magnifies the significance of algorithms in content production, social medial platform competition and digital business models constructed over such platforms. Companies, brands and influencers are thus under huge pressure to understand the unique challenges and ecosystem of emerging social media platforms (Benbya et al., 2020) and to stay ahead of this powerful digital movement (Xie et al., 2022). This special section address the timely issues associated with emerging social media platforms and deepens the understanding of the latest opportunities and challenges for digital business and consumer insights in the Pacific Asia region (Jiang et al., 2019), by offering theoretical frameworks and practical strategies that bridge academic research with industry needs
The Impact of Broadcasters\u27 Emotions and Interaction Rituals on Viewers\u27 Purchase Intention
Get PDFBackground: The rise of live streaming has transformed the e-commerce landscape, enabling real-time interaction between broadcasters and viewers while enhancing immediacy, interactivity, and emotional engagement. Therefore, studying the emotional communication and interaction methods between broadcasters and viewers holds significant importance for improving live streaming sales.
Method: The study integrated the Interaction Ritual Chain and Emotional Labor Strategy to create the combined IRC-ELS model. We proposed four hypotheses and designed two studies, which successfully collected 300 and 245 valid questionnaires respectively. Three of the hypotheses received empirical support.
Results: The broadcasters\u27 emotional labor strategy significantly influence viewers\u27 consumption intention. Viewers\u27 emotional identification with the broadcaster mediates the effect of emotional labor strategy on consumption intention. Emotional identification also directly affects consumption intention. However, emotional interaction does not moderate the relationship between emotional labor strategy and emotional identification.
Conclusion: Through the IRC-ELS model, our research enriches the literature on broadcaster behavior and viewer consumption intention. It highlights the importance of broadcasters\u27 emotional communication. The findings can also provide Asian live streaming enterprises with applicable research results on broadcaster behavior, demonstrating universal relevance