engleski

Implications of AI regulation on healthcare at the level of EU law

In this presentation critical steps toward AI regulation at the level of EU law are explored. The analysis is made in the context of upcoming regulatory proposals from the European Commission and by taking into account selected key reports and results of the recently closed consultations on the Commission’s White Paper on Artificial Intelligence. It is beyond doubt that the AI has astounding potential to transfigure healthcare by creating many breakthrough opportunities and benefits. At the same time, the legal and ethical issues raised by its deployment in healthcare are diverse and complex. In order to minimize future risks it is necessary to analyze and evaluate how it can be ensured that the AI solutions are suitable for the intended use, i.e., that they do what we want them to do, and that they do it correctly and safely. The capacity of the AI to perform tasks which are now performed by humans will lead to the replacement of human action in a number of important segments of the provision of the health services. While AI can be faster, cheaper, more accurate and advanced, it is imperative that the potential risks and negative consequences in relation to its deployment are kept in mind. Consequently, in-depth understanding of the implications that the use of AI has and may have on patients, the healthcare sector, and the society in general should be permanently cultivated. A number of concerns over deployment of AI relate to its transparency, explicability, security, reproducibility, and interpretability. For example, questions are raised whether it is possible to discover why and how AI has made a specific decision, or why and how a robot acted in the way it did. This is crucial in particular in the context of safety-critical systems that may have direct consequences for physical harm, such as medical diagnosis systems. Where regulatory responses are concerned, for example, the rules fostering principles of transparency and fairness that are already in force at the level of EU law appear to respond well to some of those risks. In other aspects, new and/or updated regulatory responses may be necessary. Overall, it is concluded that the area of healthcare necessitates appreciation of the AI as an important tool that should be designed and used in a manner serving the benefit of humankind, and that any consistent regulation at the level of EU law should primarily support that higher goal.

health data ; biometric data ; genetic data ; White Paper on Artificial Intelligence ; AI regulation

nije evidentirano