CSIRO report calls for people-centred AI in health

David Hansen and his team of scientists at CSIRO’s Australian e-Health Research Centre (AEHRC) have updated a major scientific report on AI in the healthcare system, but he likes to talk in personal anecdotes.

“I was talking to a clinician recently who does breast cancer screening, and he’s now got an AI bot which does a screen of the image as well,” Hansen tells Cosmos.

“He used to be so worried he’d miss something, he’d check something 4 or 5 times, and it was just a burden. And now he’s got this AI bot and he actually trusts it to do the second look, so he only looks twice now and he gets an opinion from the AI as well.”

It’s how Hansen, who is CEO and research director of the AEHRC, explains the valuable place AI has in health care.

He believes AI is changing the face of the healthcare system, reducing costs and improving clinical outcomes.

The report – an update to a CSIRO study from 2020 – calls on the health system to put people at the centre of AI initiatives. The document goes as far as including a primer, with case studies, on what AI is and how it will impact everyone’s lives.

But the report also points out ethical and clinical implications.

“We’re at the cusp of an extraordinary era in medicine,” Hansen says. “For the first time, machines can provide efficient administrative support for clinicians and education for patients; support diagnoses and prediction of disease and inform clinical decision making.

“The use of AI in healthcare is unique because the accuracy of models could mean the difference between life or death, or ongoing health or illness.

“A lot of people in the community – including in health – still think that AI is new thing, whereas it’s got a long history, and an impressive history. I hope it adds to the community feeling comfortable with its introduction and that we are looking at the issues around making sure it’s used properly.”

Recent breaches of privacy at the telecom provider Optus and the health insurer Medibank raised issues around data privacy and security. Hansen is hoping the report starts to change these attitudes.

“I think the paradigm shift is happening,” he says.

“I know pathology results aren’t put up straight away to make sure the doctor can get to you first. There are doctors who think that’s right and doctors who don’t, and patients who think it’s right and some don’t.

“Making sure we’re supporting people where they feel comfortable is important and so I think the most important thing is to be upfront about how people’s data is being used.

“There is a bunch of ways AI will make the health system more efficient and that frees up money to spend on patients.”

There has been a major shift in the application of AI since the 2020 report, particularly through large language models.

“We … need to ensure our nation’s preparedness for the unintended consequences of AI in the health sector,” the update says.

“We have identified trends in digital health – each being driven by the increasing digitisation of society and the increased willingness of doctors, researchers, and patients to interact using digital tools. All support the use of AI and machine learning in everyday life, including healthcare.

“To fully harness AI and machine learning (ML) we need not to just let it happen, but rather plan for its introduction into healthcare.

“This means we will be able to benefit properly from AI by ensuring the frameworks are firmly in place for ethical implementation and that the safety, quality and monitoring guidelines are established as we strive to create newer and better AI based digital tools.”

Hansen is also an advocate for a “sovereign model” of healthcare AI – one developed for the sole benefit of Australia, though the report says there are gaps in present safeguards.

“It is crucial to ensure the design, development and deployment of AI systems in Australia in legitimate, but high-risk settings, is safe and can be relied upon, while ensuring the use of AI in low-risk settings can continue to flourish largely unimpeded,” says the new report.

Work is also underway to increase the safety of AI systems in health. The Department of Industry, Science and Resources recently undertook community consultation on its Safe and Responsible AI discussion paper and is finalising decisions based on its findings.

Buy cosmos print magazine

Please login to favourite this article.