Beyond Value Elicitation: Towards Moral Profiles in Early Requirements Engineering via Role-Playing Games and Anthropologist LLMs
This study presents a proof of concept for eliciting and representing the moral profiles of digital system users in Requirements Engineering (RE) by combining immersive role-playing games (RPGs) with large language model (LLM) analysis. While existing approaches rely on predefined value taxonomies and explicit articulation, values are often tacit, context-dependent, and difficult to express directly. To address these limitations, we propose moving from the elicitation of discrete moral values to the narrative reconstruction and representation of users' moral profiles. Grounded in phenomenological and narrative anthropology, the approach focuses on capturing users' moral orientations as they emerge through situated decision-making. RPG sessions generate context-rich narrative data, which are then analyzed by a specialized LLM (GPT-A) to produce individual anthropological moral profiles (IAMPs). A validation process based on cross-comparison between model outputs and participants' responses in unseen moral scenarios assesses the adequacy of the generated representations. Results indicate that RPG environments effectively support the generation of rich, context-dependent data for eliciting tacit values, and that an anthropologically grounded LLM can transform such data into coherent narrative representations of users' moral profiles. These representations enable the contextual interpretation of users' preferences and values within the given domain, with improved performance when interpretive framing captures relationships between actions, underlying motivations, and individual domain expertise. From an RE perspective, this approach enables the analysis of user preferences and trade-offs while preserving their situated and dynamic nature, providing a foundation for integrating human moral values into the early stages of RE.
Code (0)
등록된 구현이 없습니다.
Tasks
Moral ScenariosSimilar Papers 제목 키워드 기반
On The Stability of Moral Preferences: A Problem with Computational Elicitation Methods
Preference elicitation frameworks feature heavily in the research on participatory ethical AI tools and provide a viable mechanism to enquire and incorporate the moral values of various stakeholders. As part of the elici…
Unsupervised Elicitation of Moral Values from Language Models
As AI systems become pervasive, grounding their behavior in human values is critical. Prior work suggests that language models (LMs) exhibit limited inherent moral reasoning, leading to calls for explicit moral teaching.…
On the Pros and Cons of Active Learning for Moral Preference Elicitation
Computational preference elicitation methods are tools used to learn people's preferences quantitatively in a given context. Recent works on preference elicitation advocate for active learning as an efficient method to i…
Active LearningParticipatory Moral AI Is Not Neutral: The Invisible Hand of Developers
As AI systems make more morally loaded decisions across society, one response has been moral preference elicitation. In this approach, researchers poll participants on hypothetical dilemmas and use the aggregated votes t…
The Pluralistic Moral Gap: Understanding Judgment and Value Differences between Humans and Large Language Models
People increasingly rely on Large Language Models (LLMs) for moral advice, which may influence humans' decisions. Yet, little is known about how closely LLMs align with human moral judgments. To address this, we introduc…