Abschlussarbeiten

AI companions, generative conversational chatbots, avatars, and related systems designed or used to simulate human-like relationships, are increasingly adopted as emotional support tools, wellness coaches, study partners, language tutors, customer engagement agents, therapeutic assistants, and digitally mediated confidants. Examples range from Replika, Character.AI, and Snapchat's My AI to Celebrity AI and platforms like Fanvue. Unlike conventional productivity tools, these systems derive much of their value from sustained interaction, memory, personalisation, and simulated responsiveness over time. Their effects therefore unfold relationally rather than through discrete task outputs, making constructs such as trust, dependency, wellbeing, identity, and power analytically central to their design, use, and governance.

For Information Systems, this development invites engagement with the relational aspects of emerging digital technologies. The outcomes of AI companions are not properties of the technology alone; they are co-produced through interface design, data architectures, platform incentives, governance arrangements, user practices, and broader social contexts. Early evidence is sharply double-edged: these systems may broaden access to personalised support while also enabling surveillance, commodified intimacy, behavioural influence, and new forms of dependency. This thesis can engage any facet of this landscape, individual, organisational, or societal, provided it foregrounds a clear companionship component and makes an IS contribution.

Possible Methodology:

The topic supports a wide range of designs. Candidates are encouraged to propose an approach suited to their research question, for example:

• Netnography / community analysis: Analyse how users discuss these relationships in public online communities (e.g., Reddit, Discord), following established netnographic principles (Kozinets, 2015).
• Scraped / archival data: Systematically collect and analyse publicly available posts, threads, or reviews to trace patterns across many users at scale.
• Semi-structured interviews: Conduct in-depth interviews into the formation, evolution, and impact of these relationships.
• Experience sampling: Capture real-time data on users' experiences and emotional responses during interactions.

Note: These approaches can be combined, and other methodological designs may also be viable.

Possible Research Questions:

Candidates may pursue one or more directions, for example:

1. What motivates users to adopt AI companions, and how do these relationships form and evolve over time?
2. How do design and personalisation choices (memory, voice, response style, monetisation) shape the nature and value of human–AI companion relationships?
3. What forms of trust, dependency, and power asymmetry emerge between users and platform providers?
4. How do AI companions affect users' wellbeing, identity, and social connection, positively and negatively?
5. How are emotionally sensitive relational data governed, around consent, privacy, and portability?
6. How do relational AI systems reshape practices and boundaries in organisational or service contexts (e.g., care, education, customer engagement)?
7. What tensions or paradoxes characterise sustained human–AI relationships, and how are they experienced and managed?

Candidate Requirements & Contact:

We invite candidates with an interest in qualitative or mixed-method research and in the design, governance, and use of sociotechnical human–AI systems. Curiosity about digital media, social psychology, human–computer interaction, or digital platforms is welcome. Familiarity with methods such as netnography, interviewing, content analysis, or experience sampling is a plus, but a willingness to learn the appropriate method for the chosen question matters more than prior mastery.

Potential supervisor:
Georg Voronin

Selected References:

• Ciriello, R., Chen, A. Y., Rubinsztein, Z., Vaast, E., & Hannon, O. (2025). AI, all too human AI: Navigating the companionship/alienation dialectic. Proceedings of the European Conference on Information Systems.
• Fröbel, L., & Kenning, P. (2025). What motivates people to prefer and love chatbots? An empirical investigation based on self-determination theory. Proceedings of the Hawaii International Conference on System Sciences.
• Giles, D. C. (2002). Parasocial interaction: A review of the literature and a model for future research. Media Psychology, 4(3), 279–305.
• Hoffner, C. A., & Bond, B. J. (2022). Parasocial relationships, social media, & well-being. Current Opinion in Psychology, 45, 101306.
• Horton, D., & Wohl, R. R. (1956). Mass communication and para-social interaction: Observations on intimacy at a distance. Psychiatry, 19(3), 215–229.
• Hu, D., Lan, Y., Yan, H., & Chen, C. W. (2025). What makes you attached to social companion AI? A two-stage exploratory mixed-method study. International Journal of Information Management, 83, 102890.
• Kozinets, R. V. (2015). Netnography: Redefined. Sage.
• Song, X., Xu, B., & Zhao, Z. (2022). Can people experience romantic love for artificial intelligence? An empirical study of intelligent assistants. Information & Management, 59(2), 103595.

Description

Affective polarization is defined as a growing emotional division between groups and has emerged as a defining challenge of the digital age. Social media platforms, such as TikTok, Reddit, YouTube, X or Bluesky , facilitate rapid information diffusion and enables users to form and reinforce emotion-driven communities. Due to the platform infrastructures and narratives and connected emotions propagate differently across platform ecosystems, creating reinforcement of affective issues within a digital social sphere. 

This thesis explores a cross channel analysis to understand how affective polarization manifests, evolves, and diffuses across distinct social media environments. Drawing on existing theories of networked communication and platform-mediated discourse, it examines how content characteristics (AI generated and human generated), audiences or algorithmic factors shape the emotional gap between opposing groups. By integrating data from multiple platforms, the qualitative or quantitave analysis seeks to uncover how cross-channel dynamics intensify affective divides and discovers potential pathways for mitigation of negative effects.

Requirements 

You should be interested in cognitive emotion science and how support social cohesion to connect the phenomenon of affective polarization with the information systems research. 

Potential supervisor:
Vivian Mantz

References

Bakker, B. N., & Lelkes, Y. (2024). Putting the affect into affective polarisation. Cognition and Emotion, 38(4), 418–436.

Boxell, L., Gentzkow, M., & Shapiro, J. M. (2024). Cross-country trends in affective polarization. The Review of Economics and Statistics, 106(2), 557–565. doi.org/10.1162/rest_a_01160

Iyengar, S., Sood, G., & Lelkes, Y. (2012). Affect, not ideology: A social identity perspective on polarization. Public Opinion Quarterly, 76(3), 405–431. doi.org/10.1093/poq/nfs038

Iyengar, S., Lelkes, Y., Levendusky, M., Malhotra, N., & Westwood, S. J. (2019). The origins and consequences of affective polarization in the United States. Annual Review of Political Science, 22(1), 129-146. https://doi.org/10.1146/annurev-polisci-051117-073034

Piccardi T., Saveski M., Jia C.,  Hancock J., Tsai J., & Bernstein M. (2024). Social media algorithms can shape affective polarization via exposure to antidemocratic attitudes and partisan animosity. Computers and Society.

Stieglitz, S., & Dang-Xuan, L. (2013). Emotions and information diffusion in social media—Sentiment of microblogs and sharing behavior. Journal of Management Information Systems, 29(4), 217–248. doi.org/10.2753/MIS0742-1222290408

Torcal, M., & Harteveld, E. (Eds.). (2023). Handbook of affective polarization. Edward Elgar Open Access (CC-BY-NC-ND license).

Description

The rapid spread of Generative Artificial Intelligence (GenAI) technologies, such as large language models and conversational agents, transforms how public administrations work and how citizens interact with the state and its authorities (Yun et al., 2024). New technologies such as GenAI do not only promise efficiency gains and cost reductions, they also change these interactions (Lindgren et al., 2019). Tools like GenAI-based chatbots or decision-support systems are increasingly used to answer citizen inquiries, process support cases, and to perform predictive tasks such as in policing. At the same time, delegating tasks to GenAI systems raises concerns about transparency, accountability, digital sovereignty, and algorithmic bias. In this context, public administrations become part of an AI delegation ecosystem in which authorities, citizens, and companies delegate tasks to agentic IS artifacts while also shaping these technologies through data provision, procurement, and design choices (Baird & Maruping 2021). One possible angle for this thesis could be to investigate which factors influence the adoption and enactment of GenAI in public organizations, and how this affects citizen trust, perceived legitimacy, and digital sovereignty, drawing on the IS delegation framework for agentic IS artifacts (Baird & Maruping, 2021).

The thesis may take one of the following perspectives (or even take another perspective of your choice, interest and access):

Interaction perspective: How does delegation to GenAI influence how public organizations interact with citizens, and how does this reshape these interactions?​​

Outcome perspective: How does the delegation of communication or decision-support tasks to GenAI systems in public administration influence digital sovereignty, trust, and algorithmic bias?

Depending on your interests, the thesis can focus on a specific domain (health, tax, …) or a specific level of government (e.g., municipal, state, or federal agencies). Methodologically, the thesis can apply a structured literature review (Bachelor theses only), surveys of citizens interacting with GenAI-based public service tools, interviews with public administration employees, or design-science approaches prototyping and evaluating a GenAI-based assistant.

Requirements & Contact

Interest and previous experiences with the research topic as well as familiarity with any appropriate methods and context are beneficial but not mandatory. 

Potential supervisor:
Kai Schewina

References

Baird, A., & Maruping, L. M. (2021). The next generation of research on IS use: A theoretical framework of delegation to and from agentic IS artifacts. MIS Quarterly, 45(1), 315–341.​

Kuss, P., & Meske, C. (2025). From entity to relation? Agency in the era of artificial intelligence. Communications of the Association for Information Systems, forthcoming.​

Lindgren, I., Madsen, C. Ø., Hofmann, S., & Melin, U. (2019). Close encounters of the digital kind: A research agenda for the digitalization of public services. Government Information Quarterly, 36(3), 427–436.​

Yun, L., Yun, S., & Xue, H. (2024). Improving citizen-government interactions with generative artificial intelligence: Novel human–computer interaction strategies for policy understanding through large language models. PLOS One, 19(12), e0311410.​

Description

Generative AI is increasingly used to produce game assets, dialogue, software code, music, promotional material, and interactive content. In response, digital distribution platforms have introduced different rules and technical mechanisms for disclosing AI-generated content. Steam collects free-text descriptions of AI use and differentiates between content generated during development and content generated while a game is running. itch.io provides disclosure classifications and tags for AI-generated graphics, sound, text, dialogue, and code. Roblox similarly requires disclosures for certain forms of interactive generative AI. However, these platforms differ in what must be disclosed and how disclosures are displayed. and which information is accessible for quantitative analysis. 

This thesis investigates the prevalence, characteristics, and potential consequences of generative-AI disclosure on digital creative platforms. The project can build on an existing Steam dataset containing game characteristics, AI disclosure statements, prices, release information, genres, developers, publishers, and user-review information. Depending on technical feasibility and data availability, students may collect additional data from platforms such as itch.io or Roblox.

The study may examine how developers disclose generative-AI use, which types of content and development activities are most associated with AI, and whether disclosure practices differ between platforms, genres, release periods, or developer types. Free-text statements could also be classified using dictionaries, supervised machine learning, topic modelling, embeddings, or large language models, combined with manual validation. A further direction would be to investigate whether AI disclosures are associated with observable market or audience outcomes, such as prices, review scores, review sentiment, popularity, engagement, or release strategies. 

The study will use a quantitative research design. 

Requirements

You should be interested in generative AI, digital platforms, computational social science, and quantitative research.

Students are expected to:

• Work independently with a large and potentially imperfect platform dataset.
• Clean, document, merge, and validate data from different sources.
• Use R, Python, or comparable quantitative-analysis software.
• Independently examine the feasibility of collecting data from additional platforms.
• Observe platform terms, API conditions, data-protection requirements, and responsible web-data-collection practices.

For a cross-platform thesis, students should be prepared to independently develop or adapt data-collection procedures. 

Potential supervisor:

Kevin Kintzel 

References

Feuerriegel, S., Hartmann, J., Janiesch, C., & Zschech, P. (2024). Generative AI. Business & Information Systems Engineering, 66(1), 111–126. https://doi.org/10.1007/s12599-023-00834-7

Formosa, P., Bankins, S., Matulionyte, R., & Ghasemi, O. (2025). Can ChatGPT be an author? Generative AI creative writing assistance and perceptions of authorship, creatorship, responsibility, and disclosure. AI & Society, 40(5), 3405–3417. https://doi.org/10.1007/s00146-024-02081-0

Lim, S., & Schmälzle, R. (2024). The effect of source disclosure on evaluation of AI-generated messages. Computers in Human Behavior: Artificial Humans, 2(1), 100058. https://doi.org/10.1016/j.chbah.2024.100058

Schilke, O., & Reimann, M. (2025). The transparency dilemma: How AI disclosure erodes trust. Organizational Behavior and Human Decision Processes, 188, 104405. https://doi.org/10.1016/j.obhdp.2025.104405

itch.io. (2024). Generative AI disclosure tagging. https://itch.io/t/4309690/generative-ai-disclosure-tagging

Roblox. (2026). Games with generative AI. Roblox Creator Hub. https://create.roblox.com/docs/generative-AI

Valve. (2024, January 10). AI content on Steam. Steamworks Development. https://store.steampowered.com/news/group/4145017/view/3862463747997849618