Skip to main content Skip to main navigation


Investigation of Artificial Mental Models for Healthcare AI Systems

Sabine Janzen; Wolfgang Maaß; Prajvi Saxena
In: 19th International Conference on Design Science Research in Information Systems and Technology. International Conference on Design Science Research in Information Systems and Technology (DESRIST-2024), June 3-5, Trollhättan, Sweden, Springer, 6/2024.


In the evolving landscape of healthcare, personalized Artificial Intelligence (AI) systems are vital for patient-centered care. However, patients facing health challenges often struggle with cognitive limitations, leading to incomplete or biased data that hinders their decision making abilities. To address this issue, this research in progress explores the concept of Artificial Mental Models (AMM) within healthcare AI systems. AMMs are meta representations of patient mental models, capturing their understanding and assumptions about therapy and rehabilitation processes. We present a research design for investigating AMMs in healthcare AI systems that adopts a Design Science Research (DSR) approach consisting of four iterative phases: elicitation, individualization, action, and transfer. In the elicitation phase, discrimination-free basis models are generated through web scraping and synthetic patient data. The individualization phase fine-tunes AMMs for individual patients by incorporating diverse data sources. The action phase integrates AMMs into AI systems and evaluates their real-world impact. The transfer phase applies the resulting framework to support therapy decisions for patients with compromised decision-making abilities. This research aims to enhance therapy outcomes and patient care while advancing the understanding of mental models in healthcare.