You can also listen to this podcast on iono.fm here.
Download the free LiSTN audio app on Google Play, Apple or here.
JEREMY MAGGS: Artificial intelligence (AI) is arguably one of today’s most important technological advancements, but a leading bioethicist is warning that there is a lack of guidelines on its use for healthcare in South Africa. Professor Keymanthri Moodley is the head of medical ethics at Stellenbosch University and joins us now. Professor, why have you seen fit to raise these concerns?
KEYMANTHRI MOODLEY: Well, as you know, AI in healthcare brings with it lots of potential benefit to patients and to healthcare systems and healthcare providers. But it also has a few risks that are associated and we need to ensure that we balance the benefits and the risks, as we do with any health intervention.
For this purpose, it’s important that we have guidelines for healthcare professionals and for manufacturers of AI devices, and that we have a legal framework so that we ensure, safety of products, as we do with other medication.
Read: The AI therapist is in: Can chatbots boost mental health?
JEREMY MAGGS: What principally are the risks that are of concern to you?
KEYMANTHRI MOODLEY: So there are a number of risks to consider. Some of these go back to the data that is used to build algorithms for AI technologies. As you know, we need to collect very large volumes of data to build these algorithms. Now, the quality of the data is extremely important in healthcare because we have to ensure that we have data that is adequately representative of all the different ethnic groups in our population, of gender as well, of ages, to ensure that we have good data that we are feeding into algorithms so that we can produce good AI technologies using these algorithms. So the quality of the data we collect is important.
This means that, I’m sure you’re familiar that when you usually go to your doctor or your hospital, your clinic, you are often filling in or they’re often working with paper-based records. Now, to work in the field of AI, we need to transition to electronic health records. We need to be able to collect this data, store it and use it efficiently.
So there are a number of ethical considerations linked to the data collection, as well as getting consent from patients and so on. So there’s a huge amount of work that we have to do in this area to move towards development of electronic health records throughout our health system.
Read:
SA surgeries drop on tech, health literacy gains
The NHI could make South Africa a leader in digitised healthcare
Netcare seeks digital expansion as Covid-19 pressure eases
JEREMY MAGGS: The difficulty that your sector is facing, I imagine, is to balance this rapid take up, the implementation of AI, and all the innovations it brings with what you are referring to as the need for adequate safety checks. Maybe it’s erring too much on adopting at speed and not looking in the rearview mirror.
KEYMANTHRI MOODLEY: Absolutely. So the technological development is progressing at lightning speed, as you can see.
There’s a huge emphasis on the profound benefits that AI will bring and massive investment in the technology.
But what is being left behind is the governance and the governance refers to the guidelines and the laws that need to be in place in order to ensure that we have robust safety checks, that there are also the legal provisions that are made for liability issues that may arise in the event of harm to patients.
JEREMY MAGGS: Do you know what kind of work, if any at all, the Health Professions Council of South Africa (HPCSA) is doing in this respect?
KEYMANTHRI MOODLEY: Well, it’s not clear to anybody at this point in time, we have received no communication from the Health Professions Council of South Africa. As far as I know, we are unaware that guidelines are being developed. They might be, but we are unaware of such guideline development occurring.
We do know that there are guidelines for telehealth. These were implemented, especially because during Covid there was a large need to implement telehealth consulting in the healthcare profession. But apart from those telehealth guidelines, we don’t really have anything specific for AI technologies as yet.
JEREMY MAGGS: You talk about a concept called ethical debt. What is that?
KEYMANTHRI MOODLEY: This usually occurs when we have technology developing faster than consideration is made for the ethics. So technical debt occurs when we release technical products that are not yet completely safe and ethical debt occurs when we are not taking into account the ethical considerations when we design new technologies in healthcare.
JEREMY MAGGS: Professor Moodley, thank you very much for your insight.
