Ethical, legal issues of AI use in medicine need to be resolved: Medical school dean

Artificial intelligence (AI) and robotics have led to some advancements in medicine, such as faster and more accurate diagnostic tools.

But there are many ethical and legal issues that need to be resolved before these technologies can be implemented in the field more widely, said Professor Joseph Sung, dean of the Lee Kong Chian School of Medicine at the Nanyang Technological University, yesterday.

These issues include how to ensure patients' personal data is used in a way that respects their privacy, and who should bear responsibility for mistreatment due to use of AI-enabled tools.

"If we don't know how to use technology in the appropriate way, we will be taken over and this will lead humankind to disaster," he said at the Grand Copthorne Waterfront Hotel during the annual induction ceremony of the Academy of Medicine, Singapore (AMS).

During the ceremony, the AMS conferred the title of Honorary Fellow on President Halimah Yacob in recognition of her contributions to Singapore and the academy, where she serves as its patron.

She also witnessed the induction of 105 new fellows of the academy.

In his speech, Prof Sung noted how AI and robotics have permeated many aspects of life, including in medicine, where they have impacted clinicians and healthcare providers, the healthcare system itself and the community.

For instance, image-based AI tools for diagnosis are used in many clinical specialities, he said.

He added that these tools can tell if a growth is a malignant or benign lesion, predict the prognosis and determine whether the patient is likely to respond to chemotherapy or other treatments.

"Artificial intelligence can help us to improve efficiency and allocation of resources because they can predict the major burdens of healthcare and the outcome of individuals who have cancer, diabetes or other chronic illnesses," he said.

"This is very important for policymakers to decide where the healthcare resources should be allocated."

But despite breakthroughs in AI and robotics, the adoption of solutions based on these technologies is slow, Prof Sung noted.

Data privacy is one concern, he said, noting that more private and personal data about people's health, biology and lifestyles is being generated than ever before, but there is a lack of confidence that the data is used properly.

And there is still a lot of scepticism among patients about whether AI can or should make life-and-death decisions.

Advances in AI and machine learning may lead to algorithms and recommendations that may be beyond human comprehension, Prof Sung said.

"Without a biological explanation, and without understanding of the underlying rationale of these decisions, trust could be a problem," he added.

"Can we trust a black box? That is the machine learning in our healthcare decision. So transparency and interpretability of the machine learning system or deep learning system in healthcare are very, very important."

Prof Sung drew a comparison with the debate on self-driving cars, which are also powered by AI.

"If you take this as a model, then our healthcare decisions and our surgery, we may also be moving... from a fully doctor-driven treatment, like surgeons operating on a tumour, to a fully robot-driven decision and a robot doing surgery. There will be a lot of discussion about how to apportion the responsibility."

Prof Sung delivered his lecture, titled Artificial Intelligence In Medicine: Ethical, Legal And Social Perspectives, as the 25th Gordon Arthur Ransome Orator.

The lecture series was created in 1970 to honour Professor Gordon Arthur Ransome, a pioneer in medicine here and the late founder and first master of the AMS.

About 400 medical and dental specialists and guests attended the event in person and several hundred others over Zoom.