Abstract
The article will focus on the ethical requirements for a regulation of AI in health. More precisely: meaningful human control or human-centrich approach and huma oversight; reliability of AI thorugh safety and validation of software; transparency and explainability overcoming, whern possible, the problem of opacity; equality, fairness and nondiscrimination avoiding the problem of possible bias (related to gender, age, ethnicity); data quality/accuracy in collection and data sharing, compatible with privacy; shared responsibility; information/education promoting AI literacy; sustainability.