Interactive Robotics in Medicine and Care
© Bielefeld University / Patrick Pollmeier
Exploring Human–AI Cooperation Through the Development of Intuitive and Adaptive Agents in Care Settings
This project explores how medical professionals can intuitively share their expertise with AI systems—without requiring extensive technical knowledge. The goal is to make AI-assisted systems in healthcare more accessible, adaptive, and responsive to human needs.
To study this, a virtual simulation environment is being developed that abstracts and replicates the workflows of a clinical setting. Within this environment, humans and AI agents can interact and cooperate. This approach allows medical staff to experiment safely with AI-driven systems, observe their behavior, and influence their learning process—without disrupting real clinical operations. Beyond professional users, the simulation also enables laypeople to gain playful insights into care work and human-AI collaboration.
The project contributes to improving the usability and trustworthiness of AI systems in healthcare. Insights from this research may also benefit other fields such as robotics and interactive intelligent systems, where close human–AI cooperation is essential.
Research Goals
The project aims to investigate:
- How medical expert knowledge can be integrated into AI systems effectively and intuitively.
- Which interaction modalities and machine learning approaches best support collaboration between humans and adaptive AI agents.
- How AI agents can dynamically adapt to human expertise, preferences, and workflow patterns.
The broader goal is to make AI-assisted systems more accessible, adaptive, and trustworthy for healthcare professionals. By allowing domain experts to directly shape and co-train AI agents, the research promotes systems that better reflect real clinical expertise and human expectations.