TY - JOUR

T1 - Rechtliche Aspekte des Einsatzes von KI und Robotik in Medizin und Pflege

AU - Beck, Susanne

AU - Faber, Michelle

AU - Gerndt, Simon

N1 - Funding Information: Das diesem Beitrag zugrunde liegende Vorhaben vALID wurde mit Mitteln des Bundesministerium für Bildung und Forschung unter dem Förderkennzeichen 01GP1903B gefördert. Die Verantwortung für den Inhalt dieser Veröffentlichung liegt bei den Autor:innen.

PY - 2023/6

Y1 - 2023/6

N2 - Definition of the problem: Rapid developments in the field of artificial intelligence (AI) and robotics pose new challenges not only to ethics but also to law, especially in the field of medicine and nursing. In principle, the use of AI has the potential to facilitate, if not improve, both curative treatments and adequate handling in the context of care. Administrative tasks, the monitoring of vital functions and their parameters, and the examination of tissue samples, for example, could run autonomously. In diagnostics and therapy, such systems can support the attending physician. Intelligent beds make it possible to mobilize patients early while at the same time reducing the need for personnel. Nevertheless, the use of these systems poses legal challenges. For example, there is a risk of injury to the people involved. Unlike conventional technologies, AI “suffers” from the “black box problem”: the results generated by the systems are no longer fully predictable and comprehensible. Its use entails unknown and incalculable risks, with particular implications for civil liability and criminal responsibility. Arguments: To whom the decisions of the systems are normatively attributable is a core question of legal discourse. The obvious choice, for reasons of practicability, of attributing the behaviour of an AI system to the human being who makes the final decisions is not convincing in all cases, but often degrades the human being to a symbolic “liability servant” and imposes the risks of the technologies on the human being in a one-sided manner. Furthermore, in the field of medicine and care, questions arise regarding the approval of AI systems, since the machines continue to learn during use and thus continuously change their structural design. Since the systems require any amount of reliable data for training and later use—especially through further learning—adequate handling of personal data is also necessary with regard to data protection law in the area of care and medicine. Conclusions: It is therefore advisable to address the potential for conflict from an ethical and legal perspective at an early stage in order to prevent a potential social fear of such systems and to create a practical framework for action. Orientation towards the guiding principle of “meaningful human control” offers the potential to solve these challenges.

AB - Definition of the problem: Rapid developments in the field of artificial intelligence (AI) and robotics pose new challenges not only to ethics but also to law, especially in the field of medicine and nursing. In principle, the use of AI has the potential to facilitate, if not improve, both curative treatments and adequate handling in the context of care. Administrative tasks, the monitoring of vital functions and their parameters, and the examination of tissue samples, for example, could run autonomously. In diagnostics and therapy, such systems can support the attending physician. Intelligent beds make it possible to mobilize patients early while at the same time reducing the need for personnel. Nevertheless, the use of these systems poses legal challenges. For example, there is a risk of injury to the people involved. Unlike conventional technologies, AI “suffers” from the “black box problem”: the results generated by the systems are no longer fully predictable and comprehensible. Its use entails unknown and incalculable risks, with particular implications for civil liability and criminal responsibility. Arguments: To whom the decisions of the systems are normatively attributable is a core question of legal discourse. The obvious choice, for reasons of practicability, of attributing the behaviour of an AI system to the human being who makes the final decisions is not convincing in all cases, but often degrades the human being to a symbolic “liability servant” and imposes the risks of the technologies on the human being in a one-sided manner. Furthermore, in the field of medicine and care, questions arise regarding the approval of AI systems, since the machines continue to learn during use and thus continuously change their structural design. Since the systems require any amount of reliable data for training and later use—especially through further learning—adequate handling of personal data is also necessary with regard to data protection law in the area of care and medicine. Conclusions: It is therefore advisable to address the potential for conflict from an ethical and legal perspective at an early stage in order to prevent a potential social fear of such systems and to create a practical framework for action. Orientation towards the guiding principle of “meaningful human control” offers the potential to solve these challenges.

KW - AI

KW - Care

KW - Human in the loop

KW - Liability servant

KW - Meaningful human control

KW - Medicine

KW - Participation

KW - Responsibility diffusion

U2 - 10.1007/s00481-023-00763-9

DO - 10.1007/s00481-023-00763-9

M3 - Artikel

VL - 35

SP - 247

EP - 263

JO - Ethik in der Medizin

JF - Ethik in der Medizin

SN - 0935-7335

IS - 2

ER -