Assistance Systems for Speech Recognition in Therapy and Care

Pilot 7:
Assistance systems for speech recognition in healthcare: Realization of an AI-based demonstrator system for speech-based care documentation and development of guidelines for the sustainable implementation of AI-based assistance systems in therapeutic care.

AI or machine learning-powered speech recognition and processing have a wide range of potential applications in healthcare. These technologies could greatly change healthcare professions and diagnostic and therapeutic procedures, making human-centred adoption of these technologies a key success factor.

In this work package, two different applications of automatic speech recognition will be investigated: Firstly, a system to aid nursing and medical documentation (users: healthcare professionals) and secondly, a system to aid the therapy of logopaedic speech and language disorders (users: patients).

The common goal that hsg and BO-I-T are pursuing in this pilot application is to tap the great potential of machine learning-based speech recognition and processing in healthcare. Since these technologies will significantly change healthcare professions, diagnostic and therapeutic procedures, technical and work design changes will be examined in parallel. BO-I-T aims to achieve the shared objective through a technical approach, whereas hsg concentrates on the aspects of work design and collaborates with users to implement the plan through its network.

Method development:

• Procedure model for federated learning with calculation on the user’s end device (IKA)
• Job Change Acceptance Toolbox for increasing the acceptance of AI-related change processes (WiPsy)
• Participation-oriented process model for proactive work design (IAQ)