Bachelor's Seminar
SM 452 for Bachelor's program (Business Informatics)
General Information
| Lecturer | Prof. Dr. Armin Heinzl |
|---|---|
| Course Format | Seminar |
| Credit Points | 5 ECTS (WI after Fall 2013), 4 ECTS (WI before Fall 2013) |
| Language | English |
| Grading | Seminar paper (70%), presentation (20%), discussion (10%) |
| Exam Date | See course information below |
| Information for Students | Registration: Please see information below! |
Mechthild Pieper
Contact person for Bachelor's Seminar
For further information please contact Mechthild Pieper.
Course Information
New Frontiers in Digital Transformation
Brief Description
Digital technologies and the ever-growing amounts of data are radically reshaping our daily lives as well as the economy. Embedded at the very core of the products, operations, and strategies of many organizations, digital technologies are rapidly transforming existing businesses throughout all industries. New market offerings, business processes, as well as business models are emerging around the use of these digital technologies, yielding digital innovation1. The pervasive nature of digital technology is fundamentally transforming our understanding of information systems (IS), especially regarding their development, coordination, use, and the way we interact with them. At our chair, we offer a wide range of research topics in IS, focusing on new digital technologies such as artificial intelligence (AI) and machine learning (ML). In our research, we take human-computer interaction, system design, value creation or organizational perspectives.
In our seminar, we will examine the design of digital technologies as well as their impact on individuals and organizations. In doing so, we link the offered topics to our ongoing research, which has been and is currently being published at leading international outlets.
1. Nambisan, S., Lyytinen, K. & Yoo, Y. Handbook of Digital Innovation. 2–12 (2020) doi:10.4337/9781788119986.00008.
Interested in learning more about digital innovation? Feel free to have a look at our master course IS 607 (https://www.bwl.uni-mannheim.de/en/heinzl/teaching/digital-innovation/) and/
or see Nambisan, S., Lyytinen, K. & Yoo, Y. Handbook of Digital Innovation, (2020), doi:10.4337/9781788119986. Objectives of the Seminar
In this seminar, you will acquire the ability to identify, classify, and evaluate existing research. You will learn how to develop your own research agenda as well as to present and discuss it with the participants of the seminar. You will be taught different techniques of scientific work and writing so that you will be prepared in the best possible way for the conception and writing of your Bachelor's thesis. We offer various different topic areas, which hopefully raise your interest.
Fall 2026
Registration
You may register via our online registration tool (accessible inside the university network or via VPN). During the registration period, you can select the seminar in the registration form (under ‘Application Details’ à ‘Purpose’)
Registration period: see schedule
Requirements:
- Short informal letter of motivation (maximum 1 page):
Please select a topic and give reasons for your choice, i.e., what are you particularly interested in and what do you want to learn. Please also provide two alternative topics.
- Provide your CV and your transcript of records.
We will not consider applications via e-mail or with incomplete data in the registration tool.
Topics (Fall Semester 2026)
Students are asked to write a short letter of motivation (maximum 1 page) to choose a topic and briefly justify your choice. This letter of motivation will be considered as the key reference for seminar entry, in addition to the CV and the transcript of records.
Topic Description Supervisor Emotion Data Analysis This seminar investigates users' emotional responses, measured through facial expression analysis, in interactions with a life-sized avatar. The work consists of a structured literature review on emotions in human–AI interaction, grounded in Information Systems and Human-Computer Interaction, and an empirical analysis of a dataset in which users interact with two life-sized avatars differing in appearance. Students are expected to synthesize and critically evaluate existing research, analyze differences in emotional reactions between the two avatar appearances, and explain where these differences originate. Bastian Schieck Agentic Creativity This seminar investigates the influence of AI agents and embodied avatars on human creativity, with a particular emphasis on whether such interactions inhibit or disinhibit creative performance and whether embodiment changes these effects and why. The work consists of a structured literature review grounded in Information Systems and Human-Computer Interaction. Students are expected to synthesize and critically evaluate existing research, identify gaps in the literature, and, building on these insights, design an experiment to test the effects of embodied agents on human creativity in collaboration. Bastian Schieck Data Ecosystems The growing integration and use of Generative AI (GenAI) in enterprise environments raises the demand for sharing domain-specific data that resides outside firm boundaries. Inter-organizational data sharing may become critical for realizing complex AI use cases. However, exchanging data across organizational boundaries introduces challenges regarding data sovereignty, intellectual property risks and potential loss of competitive advantage. While federated and regulatory initiatives (e.g., EU Data Spaces, EU Data Act, EU AI Act) aim at facilitating inter-organizational data exchange, their applicability, governance and scope require deeper investigation.[PARAGRAPH-BREAK]In a seminar thesis within this topic area, students are expected to systematically review contemporary research on data ecosystems and regulatory initiatives and develop a conceptual study design to examine data sharing practices within predefined domains. Dr. André Halckenhäußer Case-Based Prediction, Metric Learning, Multimodal AI Retrieval-based classifiers can make prior cases the basis of a prediction: a query is compared to its nearest labelled neighbours, and their similarity determines the outcome. This promises explanations that are exact by construction, yet it raises technical questions. How should a similarity metric be learned so that retrieved cases remain recognizable to humans, how can image, text, and cross-modal interactions contribute separately to a comparison, and what predictive cost do such constraints impose compared to unconstrained deep fusion? This thesis should review metric learning and retrieval-based prediction approaches for multimodal data (e.g., frozen encoder features, learned projections, bilinear interaction models). Students should implement a small prototype comparing a retrieval-based classifier against a standard classifier on a multimodal benchmark, analysing the accuracy-interpretability trade-off, or conduct/ plan an experiment based on an existing solution. Solid coding skills are required. Florian Rüffer Case-Based Explanations, Retrieval-Based Prediction, XAI Explaining AI decisions through similar prior cases with known outcomes aligns with how humans naturally reason: by analogy and comparison rather than through abstract feature weights. However, how case-based explanations and their attributes actually affect human reasoning remains an open question. Which properties of presented cases (e.g., similarity, number, contrast between outcomes, richness of detail) help users assess AI advice, and which merely increase persuasion and over-reliance? This thesis should review empirical research on how case-based explanations influence human judgement, trust, and decision performance in AI-assisted settings. Students should synthesize findings into a framework of explanation attributes and their cognitive effects or conduct/ plan a small experiment on one selected attribute. Basic knowledge of empirical methods is advantageous. Florian Rüffer Synthetic Data (focus healthcare) The growing use of generative AI and advanced simulation techniques has created new opportunities to generate synthetic data that can resemble real-world observations without directly relying on data collected from humans. Such synthetic data, including human surrogates, human digital twins, digital humans or simulated populations, can help address challenges related to data scarcity, privacy, access restrictions, and ethical concerns. At the same time, questions remain regarding how accurately synthetic representations capture relevant characteristics of real-world populations and whether they can reliably serve research purposes.[PARAGRAPH-BREAK]For this seminar thesis, students are expected to synthesize the current state of knowledge on synthetic data and related concepts, examining their generation, evaluation, and application. A particular focus on healthcare is conceivable, for example, investigating the use of synthetic patient data, virtual patients, or human digital twins for medical research, clinical decision support, or healthcare simulations. Based on a selected use case, students should generate a synthetic dataset (or develop a coherent approach for its generation), and critically assess its quality, applicability, and limitations. Mechthild Pieper Synthetic Data (focus creativity) The growing use of generative AI and advanced simulation techniques has created new opportunities to generate synthetic data that can resemble real-world observations without directly relying on data collected from humans. Such synthetic data, including human surrogates, human digital twins, digital humans or simulated populations, can help address challenges related to data scarcity, privacy, access restrictions, and ethical concerns. At the same time, questions remain regarding how accurately synthetic representations capture relevant characteristics of real-world populations and whether they can reliably serve research purposes.[PARAGRAPH-BREAK]For this seminar thesis, students are expected to synthesize the current state of knowledge on synthetic data and related concepts, examining their generation, evaluation, and application. A particular focus on human-AI creativity is conceivable. Since we have a large dataset of human-AI image generation interaction traces, the student could attempt to create a synthetic dataset from the given dataset to use it for further research, and critically assess its quality, applicability, and limitations. Deborah Mateja Virtual Try On Technologies Virtual try-on (VTO) systems are technologies that allow users to virtually experience products on themselves before purchasing them, and are increasingly used in e-commerce and digital retail. Different technological approaches have been developed to enable virtual try-on, ranging from augmented reality applications that superimpose products onto images or live video, to avatar-based systems that digitally represent the user, and more recently, generative AI approaches that generate realistic images of users wearing selected products. These approaches differ substantially in how users and products are represented, which data and technological capabilities they require, and the types of products and applications for which they can be used.[PARAGRAPH-BREAK]In order to understand the current technological possibilities and future development of virtual try-on, it is important to understand the different technologies and algorithms through which VTO can be implemented. For this seminar paper, the student is expected to conduct a structured literature review to identify and classify the technologies, methods, and algorithms that have been used in previous studies to enable virtual try-on, and to compare their requirements, capabilities, limitations, and areas of application. Ideally, the student will identify or generate illustrative examples of the different approaches using publicly available algorithms, models, or VTO technologies to demonstrate how these technological approaches are instantiated in practice and empirically evaluate the technologies using a small explorative sample. Rosa Holtzwart Hospital AI Platform Governance At the research campus M2OLIE at the University Hospital Mannheim, diagnostic and treatment cycles are being systematically shortened through closed-loop, technology-supported care processes. Supporting the clinical decision processes within these cycles increasingly points toward an AI platform — which raises a question that remains poorly understood: how are AI platforms in healthcare organizations actually governed? This seminar thesis examines which governance mechanisms are described for AI platforms in healthcare organizations, and how responsibility for these mechanisms is allocated between platform provider and deploying organization. The expected outcome is a structured overview of governance mechanisms for AI platforms in healthcare, including an assessment of their coverage, quality and need for further investigation. The work thereby contributes to the discussion of platform governance under institutional constraints and, in practical terms, to the design and management of the AI platform envisaged at M2OLIE. Michael Sternberg Course Outline & Schedule (preliminary)
Event Time Period / Deadline Deliverables Registration Period 24.08.2026 – 04.09.2026 Registration via the online tool – Attach your CV, transcript, and motivation letter Sending of Confirmations 07.09.2026 (end of day) Withdrawal Deadline 08.09.2026 (end of day) Kick-Off Workshop 15.09.2026 Participation in the kick-off introductory event
Contact and meeting with your supervisor1st Milestone 29.09.2026 Submit first draft to your supervisor: Detailed outline – Bibliography 2nd Milestone 27.10.2026 Submit second draft to your supervisor: Table of contents – Introduction: fully formulated – Methodology: fully formulated – Results: structured draft – Discussion: structured draft Submission of Paper 10.11.2026 (noon) The seminar paper must be submitted in digital and printed form on the submission day. Send the PDF version by 12:00 pm at the latest via email to Mechthild Pieper (pieper@uni-mannheim.de) and CC the chair's secretariat (wifo1@uni-mannheim.de) as well as your *supervisor*. Additionally, two printed copies must be submitted to the secretariat on the same day. The submission is only considered complete if all three steps are completed on time. Submission of Presentation 15.11.2026 (noon) Optional: Ask your supervisor for feedback on the presentation in advance – Send your presentation in PDF format via email to Mechthild Pieper Presentation 17.11.2026 Attend the seminar and actively participate in the discussion on the seminar day – Present and discuss your seminar paper in the joint workshop – Discussion and feedback for at least one seminar paper of other students Literature
- Webster, J., & Watson, R. (2002). Analyzing the Past to Prepare for the Future: Writing a Literature Review. MIS Q., 26.
- Leidner, Dorothy E. (2018) “Review and Theory Symbiosis: An Introspective Retrospective,” Journal of the Association for Information Systems: Vol. 19 : Iss. 6 , Article 1.
To access the literature you have to be in the VPN of the University of Mannheim.