Ethical issues around artificial intelligence (AI) must not be outsourced to tech companies, Dr Adrian Brady, Consultant Radiologist, Mercy University Hospital, Cork, told a recent conference in Dublin.
Speaking on the ethics of AI at the Health Informatics Society of Ireland annual conference 2019, Dr Brady raised a number of challenges around the application of AI in radiology.
“Many of the people in my specialty and my profession are very concerned that the work we do today may in the future no longer be there for us. And while I don’t really believe that is the case, I believe a lot of our work will change.”
Dr Brady added that “human doctors go through a very prolonged period of training, to become human doctors… [it is] very important that we provide the influences to ‘Doctor Google’ to make sure that Doctor Google operates properly and in a human way.”
He said it will be important to have transparency on how AI decisions are made.
“Of course this isn’t very attractive to some commercial developers, because of the whole transparency of it, the more liable it is to
danger of hacking or someone stealing intellectual property.”
Dr Brady also said “liability” has to be discussed.
“If something goes wrong with an AI radiological application, is it the radiologist who is involved in the interpretation who is liable? Or the manufacturer, or the software developer? Or the hospital? I don’t know, but it is something we have to think about.
“From a radiological perspective, we need to develop a code of ethics and practice and update them…we must not outsource our ethical thinking to the tech companies.”