Misleading AI-generated doctors pose ‘huge danger to public safety’