Generative AI in IS Research and Education: Opportunities and Challenges

Permanent URI for this collectionhttps://hdl.handle.net/10125/112560

Browse

Recent Submissions

Now showing 1 - 4 of 4
  • Thumbnail Image
    Item type: Item ,
    AI-Enhanced Literature Reviews: Connecting Emerging Phenomena and Bodies of Knowledge
    (2026-01-06) Naqvi, Syed Asad Ali; Zimmer, Markus; Kauschinger, Martin; Drews, Paul; Basole, Rahul
    Bodies of Knowledge (BoKs) are structured knowledge collections that describe key concepts, terminology, and practices. Emerging research phenomena are rapidly evolving and novel domains that are attracting growing interest from researchers. The relationship between existing bodies of knowledge and emerging phenomena is complex, and offers exciting research opportunities. To address this challenge, we introduce a literature review framework called Contextualize, Visualize, and Interpret (CVI-AI), which leverages artificial intelligence and visualization capabilities to help researchers understand the relationship between emerging phenomena and existing BoKs. Our contributions are threefold. First, we illustrate how incorporating AI tools into the CVI-AI framework improves the literature review process and outcomes by examining the connections and semantics within the literature while formulating novel research questions. Second, our framework provides guidance for applying AI tools to research. Third, our study supports researchers in developing new research agendas by linking emerging phenomena to the existing BoKs.
  • Thumbnail Image
    Item type: Item ,
    Prompt Engineering as a Cognitive Interface: Reframing Human-AI Collaboration and Digital Literacy in the Age of Generative AI
    (2026-01-06) Azizian, Sasan; Mehrabian, Reza; Honary, Vahraz
    The rapid evolution of generative artificial intelligence (AI), typified by systems such as GPT-4, Claude, and DALL-E, has introduced new paradigms for human-machine interaction. At the center of this transformation lies the practice of prompt engineering, the act of crafting natural language instructions to elicit specific, high-quality responses from AI models. Traditionally regarded as a technical workaround or hack, prompt engineering is increasingly recognized as a structured, strategic, and cognitively rich process. This paper reframes prompt engineering as a cognitive interface, a conceptual bridge through which humans externalize mental models, intentions, and iterative hypotheses into prompts that shape AI behavior. Drawing on a mixed-methods study that analyzed 300 real-world prompts and 15 interviews with expert prompt designers, we introduce the Prompt Cognition Loop (PCL). This novel framework describes prompting as a four-phase cycle: 1) mental modeling, 2) semantic projection, 3) dialogic feedback, and 4) intent refinement. This model aligns with key theories in cognitive science and human-computer interaction, including schema theory, mental models, and reflective practice. We further explore the pedagogical potential of prompt engineering in digital education, suggesting curricular and design strategies for developing prompt literacy.
  • Thumbnail Image
    Item type: Item ,
    Adaptive Kernels in DCGANs
    (2026-01-06) Khandaker, Shamir; Ahmad, Sahan; Islam, Aminul
    Deep Convolutional Generative Adversarial Networks (DCGANs) is a specialized deep learning architecture tailored for image generation tasks. DCGANs have notably enhanced the field of image generation by introducing robust training techniques and architectural principles that yield high-fidelity images that closely mirror those present in the training dataset. Recent research suggests that employing a deterministic adaptive kernel postconvolution can bolster CNN’s generalization capabilities, thereby benefiting DCGANs. However, the challenge of generating the weights for these deterministic adaptive kernels remains an active area of research. In response, we propose an innovative adaptive kernel method that utilizes the convoluted data from a layer to generate a dynamic set of four Gaussian kernels. Subsequently, convolution operations are performed on the convoluted data using either a single Gaussian kernel at a time or all four kernels sequentially, with the order of application rotating after each epoch. To validate our novel adaptive kernel approach, we conducted experiments using DCGAN and one of its variants, evaluating performance on two diverse datasets (CIFAR-10 and CIFAR-100) for image generation tasks. Importantly, our method seamlessly integrates with any DCGAN variant as a plug-and-play solution, introducing no additional trainable parameters to the network. We also offer a comprehensive analysis of the impact of our proposed adaptive kernel method.
  • Thumbnail Image
    Item type: Item ,
    Introduction to the Minitrack on Generative AI in IS Research and Education: Opportunities and Challenges
    (2026-01-06) Tahmasbi, Nargess; Shan, Guohou; Rastegari, Elham; French, Aaron