RESPONSIBLE ROBOTICS (RR-AI) Tracing ethical and social aspects of AI-based transformations in healthcare work and knowledge environments

Project description

The integration of embodied Artificial Intelligence (AI) into healthcare and society is expected to deliver major benefits in future decades. However, innovations such as AI operating robots, AI prosthetics, care- or at some point even micro- and nanorobots will come with a number of ethical, social, political and legal challenges, among them ground-breaking shifts in the work cultures and expertise of medical professionals. These challenges arising from novel divisions of labour between humans and machines need to be addressed proactively if embodied AI is to be implemented into medicine and society successfully and responsibly. While overarching principles such as those by the European High Level Expert Group on Artificial Intelligence, or standards such as the ISO for personal care robots have been developed, concrete and fine-grained frameworks for a responsible integration of embodied AI products into healthcare practices and work cultures are still largely missing. There are also no best practice models available for the interdisciplinary development of human-machine applications in biomedicine that take ethical, social and regulatory issues into account.

The joint research project Responsible Robotics therefore sought to:

empirically study the social and ethical and legal dimensions of two novel AI-based technologies – a service robot named GARMI, and a smart arm exoprosthesis – as they are being developed and implemented in healthcare practice;
develop a practical toolbox for future interdisciplinary AI innovation, as well as concrete standards and recommendations for responsible integration of embodied AI into healthcare work practice and training;
experimentally test these tools and recommendations through interdisciplinary co-creation and work-place integration of embodied AI applications.

The project thus takes an innovative “embedded” approach, whereby ethical, social, legal and political analyses constitute integral elements of an AI product design process as well as its work place integration.

Against the theoretical backdrop of the responsibility concept, the team classified four ethical and legal principles for robots in elderly care:

Data minimization by collecting only necessary data for the intended purpose.
Data security through the protection of sensitive user data.
Transparency by accurately representing the behavior of the robotic system.
Data visualization so that users can determine usage themselves.

To implement these principles, the team proposed the implementation of a data recorder with encryption, hashing, and signing techniques in the robot. The team identified the necessary legal acts for the use of robots, including the General Data Protection Regulation, the Medical Devices Regulation, and the AI Regulation.

Furthermore, the researchers propose introducing a series of best practices for the Embedded Ethics and Social Science approach. The project results have been discussed with stakeholders, tested in pilot projects, and widely disseminated.

The project was completed by 30 June, 2023.