When AI supports caregivers but doesn't replace them

Key Takeaways:

• AI supports, humans decide: Responsible AI complements subject-matter expertise but does not replace it.
• Developed collaboratively, implemented in a practical way: Users actively help design the technology, thereby increasing its acceptance and effectiveness.
• More time for what matters most: Digital assistance takes the burden off routine tasks and frees up time for working with people.

Artificial intelligence promises more efficient processes and relief from routine tasks. In the field of nursing, however, a key question arises: How can AI provide support without replacing the responsibility of healthcare professionals?

The Interreg-ABH-funded research project ADOPI (“Work Support in the Documentation of Nursing Interventions”), led by the FHV, is dedicated to addressing this question. Together with partners from Austria and Germany, the researchers are developing digital assistance systems that support nursing staff in documentation tasks while preserving their professional autonomy in decision-making.

“The real question isn’t whether AI can be used in nursing. The question is how we design AI so that it supports nursing professionals rather than replacing them,” emphasizes project leader Philipp Lizat from the FHV’s Human-Centered Technologies Research Centre.  

More time to work with people 

Documentation is essential for quality assurance and the traceability of nursing interventions, but it is very time-consuming. Research conducted as part of the project shows that 84 percent of the nursing professionals surveyed spend up to one hour a day on documentation. Added to this are various documentation systems, the parallel use of software, paper forms, and notes, as well as the ongoing updating of care plans. These challenges formed the starting point for the development of digital support solutions. The goal is to simplify processes, improve the consistency of documentation, and reduce the workload on nursing professionals.  

Trustworthy AI is developed together with users 

A key feature of the project is its human-centered approach. Nursing professionals were actively involved in the development from the very beginning. Their experiences and requirements were directly incorporated into the design of the software solutions. The care plan assistance tool developed also follows this principle. The system uses artificial intelligence methods to process information and generate suggestions. However, the decision itself always remains with the human. The researchers deliberately rely on a so-called “human-in-the-loop” architecture: The AI supports, structures, and makes recommendations, but does not assume responsibility for care decisions. “Care measures always involve people. That’s why the final decision must remain with a human,” says Philipp Lizat.

A model for the responsible use of AI 

In addition to developing new software, ADOPI provides important insights into how AI can be designed for use in sensitive areas: transparently, in a traceable manner, and under human control. The close collaboration between nursing practice, research, and software development builds trust and enables solutions that are tailored to the real needs of users.  Thus, the significance of the project extends far beyond nursing documentation. ADOPI exemplifies how trustworthy AI can be developed: not as a replacement for human expertise, but as a tool that supports people in their work. 

Contact:

Philipp LIZAT, BA
Student Research Assistant

+43 5572 792 3724
philipp.lizat@fhv.at