Goal of the research project:
Developments in Artificial Intelligence (AI) offer new innovative ways to contribute to the well-being and progress of individuals and society. However, due to a multitude of incidents with AI (e.g., discrimination through AI predictions), the accountability of AI becomes more and more important. In general, accountability means that actions performed can be clearly assigned to a person. Applied to AI, accountable AI-based information systems (AAIS) refer to a socio-technical set of relationships consisting of humans interacting with AI technologies to perform certain tasks, where the actions taken in the course of the interaction can be uniquely attributed to a person. AAIS is intended to ensure that someone can be held legally responsible if the AI- based IS fails.
While calls to develop and embed mechanisms to create accountability in AI-based IS are growing, research on AAIS is still in its infancy. Based on the current state of research, there is conceptual ambiguity about AAIS due to the multitude of definitions of terms and differing opinions on what objectives should be pursued by creating accountability. Second, there is a lack of validated evidence and explanatory models on how the use of accountability mechanisms affects the development, operation, and use of AI. This gives rise to three research questions, which this project aims to answer: (1) “What facets of accountability are relevant to AAIS?”, (2) “What impact does accountability have on perceptions during the development, operation, and use of AI?”, and (3) “How does accountability affect the behavior of AI users and architects?”.
To develop a holistic conceptualization of AAIS, the applicants plan to identify facets of accountability that are critical to the effectiveness of accountability mechanisms. This may provide a conceptual methodological toolkit for further research on accountability. The second research question will be answered by examining both positive and negative perceptions of stakeholders regarding accountability mechanisms. In doing so, the research draws on accountability facets and is theoretically grounded by the Accountability Theory. The insights gained make a theoretical contribution in the form of multi-perspective explanatory models of the impact of accountability. To answer the third research question, real-world research will be conducted to investigate how the behavior of AI architects and users changes when they can be held accountable through accountability mechanisms.
Key Data:
Contact:
- Prof. Dr. Alexander Benlian (benlian@ise.tu-…),