Department of Mathematics and Computer Science

Personas are fictional user profiles used in human-centered design to represent key characteristics (Pruitt & Grudin, 20023), such as goals, needs, and values. They are manually created through user research and are widely used in fields such as healthcare and education.

Expanded by large language models (LLMs), AI-driven personas offer a dynamic alternative (Asadi & Kropczynski, 2024). These personas are derived from large datasets (Nielsen et al., 2022; Sankalp et al., 2025; see also Salminen et al., 2020), including social media and customer service transcripts, and can be continually and efficiently evolved. However, AI personas often lack a deep understanding of the human values they are intended to express. This leads to generic or stereotypical profiles that might fail to truly represent users (Lazik et al., 2025).

This master’s thesis aims to develop a value-based prompt design approach for building AI personas that better reflect diverse human values, asking: How can we consider human values into representative AI personas? The work will be grounded in human-centered design and draw on recent research in prompt engineering and value sensitive design.

Procedure

1. Make yourself familiar with foundational research on personas and value sensitive design.
2. Conduct a literature review on prompt engineering and existing methods for generating AI personas.
3. Construct a value-based prompt design script:
   1. Identify relevant human values (e.g., Sadek et al., 2024)
   2. Translate these into structured questions (value questionnaire)
   3. Develop a systematic prompt design script to generate AI personas reflecting these values
4. Evaluate the prompt design script:
   1. Collect real data by letting participants fill out your value questionnaire through a qualitative user study
   2. Generate multiple AI personas using the developed prompts
   3. Validate the AI personas by comparing them with the value questionnaire answers
   4. Analyze how well different value orientations are represented
5. Refine the value-based prompt design script based on evaluation findings

References

• Amir Reza Asadi and Jess Kropczynski. 2024. Qualitative Data-Driven Personas: Designing an Interactive System for Creating AI Personas. In Proceedings of the 2024 The 6th World Symposium on Software Engineering (WSSE) (WSSE '24). Association for Computing Machinery, New York, NY, USA, 232–236. https://doi.org/10.1145/3698062.3698096
• Christopher Lazik, Christopher Katins, Charlotte Kauter, Jonas Jakob, Caroline Jay, Lars Grunske and Thomas Kosch. 2025. The Impostor is Among Us: Can Large Language Models Capture the Complexity of Human Personas?. arXiv preprint. https://doi.org/10.48550/arXiv.2501.04543
• Lene Nielsen, Bernard J. Jansen, Joni Salminen, José Abdelnour Nocera, and Soon-Gyo Jung. 2022. Personas: New Data, New Trends. In Extended Abstracts of the 2022 CHI Conference on Human Factors in Computing Systems (CHI EA '22). Association for Computing Machinery, New York, NY, USA, Article 139, 1–3. https://doi.org/10.1145/3491101.3503772
• Malak Sadek, Marios Constantinides, Daniele Quercia, and Celine Mougenot. 2024. Guidelines for Integrating Value Sensitive Design in Responsible AI Toolkits. In Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems (CHI '24). Association for Computing Machinery, New York, NY, USA, Article 472, 1–20. https://doi.org/10.1145/3613904.3642810
• Joni Salminen, Kathleen Guan, Lene Nielsen, Soon-gyo Jung, and Bernard J. Jansen. 2020. A Template for Data-Driven Personas: Analyzing 31 Quantitatively Oriented Persona Profiles. In Human Interface and the Management of Information. Designing Information: Thematic Area, HIMI 2020, Held as Part of the 22nd International Conference, HCII 2020, Copenhagen, Denmark, July 19–24, 2020, Proceedings, Part I. Springer-Verlag, Berlin, Heidelberg, 125–144. https://doi.org/10.1007/978-3-030-50020-7_8
• Sankalp Sethi, Joni Salminen, Danial Amin, and Bernard J Jansen. 2025. "When AI Writes Personas": Analyzing Lexical Diversity in LLM-Generated Persona Descriptions. In Proceedings of the Extended Abstracts of the CHI Conference on Human Factors in Computing Systems (CHI EA '25). Association for Computing Machinery, New York, NY, USA, Article 35, 1–8. https://doi.org/10.1145/3706599.3719712
• John Pruitt and Jonathan Grudin. 2003. Personas: practice and theory. In Proceedings of the 2003 conference on Designing for user experiences (DUX '03). Association for Computing Machinery, New York, NY, USA, 1–15. https://doi.org/10.1145/997078.997089