SINGAPORE – As generative artificial intelligence (gen AI) becomes more widely used in healthcare, the sector will need to work with the tech industry and regulators to develop guidelines for the use of this new technology.
Associate Professor Daniel Ting, director of Singapore's Ministry of Health's AI Office, said this will enable the development of innovations using next-generation AI while also ensuring its safe use.
He pointed to a variety of uses for the technology in medicine, such as quickly summarizing clinical records and other patient data and proofreading medical literature.
He said that in the longer term, Generation AI may be used by doctors to provide diagnoses and prognoses, but noted that a regulatory framework would need to be put in place before such use could be permitted in the country.
Prof Ting added that any framework governing the use of Generation AI in healthcare would need to assess several criteria, including that the responses generated are not subject to AI “hallucinations” – referring to inaccurate results generated due to factors such as a lack of training data – and are consistent with clinical consensus.
He noted that regulations often lag behind new technology, and that including regulators in the development process is important to ensure the final product is safe and compliant.
“I think the key is to start the discussions early,” he said.
Prof Ting spoke to reporters on the sidelines of a symposium titled “Planning the Next Phase: Emerging Trends in AI in the Healthcare Sector.”
The symposium, held at the Capella Singapore Hotel on Sentosa Island, was part of the Asia Tech x Singapore event, which ends on May 31.
When asked how such guidelines fit into the model AI governance framework for Generative AI being developed by the AI Verify Foundation and the Infocomm Media Development Authority, Prof Ting said the framework for gen AI in healthcare is still in the early stages of conceptualisation.
He added that input from other stakeholders, such as medical experts, would also be needed before such a framework could become a reality.
Professor Ting pointed to the many potential applications of Gen AI in healthcare and said different frameworks may be needed to monitor such diverse uses.
His presentation included research findings from a paper in the Lancet Digital Health journal published April 23.
Professor Ting is corresponding author of a study which identified several challenges in deploying large language models (LLMs) – AI systems that can process vast amounts of text to understand and generate human language – in healthcare.
These include the need for security measures to prevent the inadvertent disclosure of identifiable patient data captured during LLM training, and the need for patients to give clear informed consent for their data to be used for such purposes.
