Abstract
The use of machine learning (ML) in medicine is becoming increasingly fundamental to analyse complex problems by discovering associations among different types of information and to generate knowledge for medical decision support. Many regulatory and ethical issues should be considered. Some relevant EU provisions, such as the General Data Protection Regulation, are applicable. However, the regulatory framework for developing and marketing a new health technology implementing ML may be quite complex. Other issues include the legal liability and the attribution of negligence in case of errors. Some of the above-mentioned concerns could be, at least partially, resolved in case the ML software is classified as a ‘medical device’, a category covered by EU/national provisions. Concluding, the challenge is to understand how sustainable is the regulatory system in relation to the ML innovation and how legal procedures should be revised in order to adapt them to the current regulatory framework.
