A report by the European Academies Science Advisory Council and the Federation of European Academies of Medicine recommended that high-risk artificial intelligence systems in healthcare remain under strict human supervision.