The article “Toward a Clearer Process for Value Sensitive Artificial Intelligence” by Christina Cociancig (University of Bremen) and Professor Hendrik Heuer, head of the CAIS research program “Design of Trustworthy Artificial Intelligence”, was recently published in the journal Science and Engineering Ethics. The publication emerged from their collaboration as part of Cociancig’s doctoral dissertation.
The paper addresses the question of how ethical and societal values can be systematically integrated into the development of AI systems. Value Sensitive Design (VSD) provides an established, proactive, and iterative framework for this purpose. It comprises three types of investigations:
- Conceptual investigations: Who are the relevant stakeholders, and which values are important to them?
- Empirical investigations: What are the stakeholders’ experiences, needs, and priorities?
- Technical investigations: How can these values be concretely implemented in the functionality of AI systems?
In the article, Cociancig und Heuer analyze how this methodology is applied in the emerging field of Value Sensitive Artificial Intelligence (VSAI). They conducted a systematic literature review to examine how researchers describe and implement VSD in AI projects. The authors connect the theoretical foundations of VSD with the practical challenges of AI systems, which are often difficult to clearly delineate.
Their analysis identifies both theoretical and practical gaps:
- Theoretical: AI systems vary widely – from individual machine learning models to complex socio-technical systems. This makes it difficult to identify relevant stakeholders and values.
- Practical: Many studies reference VSD but do not consistently implement all three types of investigations in an integrated manner.
Based on these findings, the authors develop a framework that highlights these gaps and provides a synthesis of literature findings and their own recommendations. This results in concrete, actionable strategies for the development of value-sensitive AI systems. The goal is to give researchers and practitioners clear guidance and to strengthen the development of AI systems that are more strongly aligned with ethical values.
—
About the paper
Heuer, H. & Cociancig, C. (2026). Toward a Clearer Process for Value Sensitive Artificial Intelligence. Science and Engineering Ethics. https://link.springer.com/article/10.1007/s11948-026-00583-2
The article builds on the earlier work “AI ethics unwrapped: an empirical investigation of ethical principles in collaborative ideation processes”: https://link.springer.com/article/10.1007/s43681-024-00638-9
It is also part of Christina Cociancig’s dissertation “AI ethics unwrapped: understanding and shaping ethical AI through value sensitive design”: https://media.suub.uni-bremen.de/entities/publication/d390d5fc-e523-44e8-93b7-5f70325d8ad1