AI offers enormous potential benefits in healthcare, leading Cork doctor says

Artificial intelligence (AI) offers enormous potential benefits in healthcare, a leading Cork doctor has said, but human beings must be kept at the heart of the decision making process.

Professor Richard Greene of University College Cork, who is an obstetrician and gynaecologist, as well as the HSE’s chief clinical information officer, recently served on a citizens’ jury on the use of AI in healthcare.

Organised by the Irish Platform for Patient Organisations, Science, and Industry (IPPOSI), the 24-person jury convened from last September to December to offer the public’s perspective on this complex topic.

Prof Greene served as a member of the independent Jury Oversight Panel, and he spoke to The Echo about the jury’s findings.

“The citizens’ jury was set up to get a feeling from the citizens as to the potential benefits of AI in healthcare, and they were randomly selected right across the country, right across appropriate gender and ethnic membership in comparison to the composition of the population, and they were broadly in favour of the use of artificial intelligence, but they wanted a lot of controls around it,” he said.

The jury published 25 recommendations on the safe, ethical, and inclusive use of AI in Ireland’s healthcare system, and it wrote to the Government setting out the need for an independent regulator and a commissioner to oversee AI, and a five-year national strategy on AI in healthcare.

Members of the jury endorsed the early, low-risk deployment of high-quality, human-monitored AI tools in helping alleviate pressures on the healthcare system, in pioneering advances in care, and empowering people to take a more active role in their own health. However, they also emphasised the need for strong regulation, transparent oversight and robust data security.

“The jury was very clear that they wanted a regulator around the use of artificial intelligence in healthcare, and that the regulator would handle legislation to enact the EU AI Act,” Prof Greene said.

“There is a lot of activity going on in the services around this as well, the Department of Health and the HSE have committed to an AI implementation strategy, which is also in the Programme for Government, there’s also stuff come out from the Department of Enterprise around it, the secretary of Hiqa has also been asked by the minister of health to develop guidance to promote the responsible use of AI in cognisance of the EU AI Act and the ethical side of AI use.

“We ourselves in the HSE are currently in the process of drawing up a framework for adoption of AI and guidance around it, so that if people in the services are looking at it, there are approaches and frameworks there to ensure that people would be able to approach it faithfully.”

The jury’s findings clearly corresponded with prevailing opinion, he said, that where the use of AI in healthcare was concerned, human beings must be kept at the heart of the decision making process.

“Generally in this country, and internationally, there’s a belief that we need to continue to have the human in the loop with AI, in other words that it’s not making decisions on its own, that we have human monitoring to keep an eye on the output from it, and to ensure that we pick up any risks from it as well.

“That came across very strongly from the citizens’ jury, but it’s also one of the things that any of us who are interested in it are also talking about,” Prof Greene said.

He added that a number of other concerns about the future use of AI were raised by the jury.

“They wanted equitable access, that it be for the benefit of all and not just based on any disparity on public/private healthcare, and they were also very keen that there would be a whole piece around public education and engagement, like a national awareness campaign to inform the public of the benefits of it, the risks of it, and the role of it in healthcare.

“They talked a lot about consent as well, that people should have know that it was being used and could decide not to have it used in their care,” he said.

Prof Greene noted that AI was currently being used beneficially in medicine to screen mammograms, and to look for fractures in bone x-rays, examples he said of instances where AI was being used very well.

“In fact, by using it, you’re really increasing the chances that the doctor will pick up on the fracture rather than miss the fracture,” he said.

The need for a human hand on the AI wheel was something that had come very strongly in the findings of the citizens’ jury, he said, and that was a sentiment replicated in international studies.

“I don’t think people are in any way ready to have AI be their doctor or their nurse, in a situation where it’s working on its own, but rather that it’s working in support of clinicians and healthcare to do a better job,” he said. The sci-fi future of AI doctors is still some way off, Prof Greene added.

“Instead, AI will be helping doctors and nurses, helping them to make diagnoses, or perhaps suggest to them treatments for particular patients, but it won’t be making the decisions, it will still be a human being writing the prescription for the patient.”

Key recommendations of the citizens’ jury on the use of AI in healthcare in Ireland:

• There should be human oversight at all points of AI-enabled healthcare, from its development to its deployment.
• The use of AI-enabled care should be equitably accessible to all, and there should be no distinction between care accessed publicly or privately.
• The implementation of the EU AI Act should be enshrined in national legislation, and there should be special provision to protect health data relating to AI.
• The Attorney General should advise the Government on enacting additional legislation this year to close gaps in the EU AI Act, ensuring safeguards against the misuse of health data by exempt entities.
• A national strategy on the use of AI in healthcare in Ireland should be published in 2026.
• The benefits of using AI in healthcare should be returned to the public to reflect the valuable contribution made to innovation by sharing their health data.
• The budget should allocate adequate, multi-annual funding to begin investment in quality AI systems for healthcare in Ireland.
• A fully representative, multi-disciplinary statutory body should be established in 2025 to regulate and enforce standards around the use of AI in healthcare, including licensing, compliance monitoring, and imposing penalties.