AI used for treatment must have doctor oversight, says College

Artificial intelligence (AI) comes with risks as well as benefits, but must have specialist doctor oversight when treating patients, says the Royal Australian and New Zealand College of Radiologists (RANZCR).

The College has released two new position statements to guide radiologists and radiation oncologists on the use of autonomous AI and generative AI in clinical settings.

However, both position statements outline the risks of AI and the need for human supervision when using it as a clinical tool to perform tasks and process data.

RANZCR believes that when used correctly, AI can significantly reduce the burden on specialists by streamlining tasks and reducing clinical workload which can lead to enhanced patient care.

Risks associated with AI as a medical tool include issues related to confidentiality, hallucinations, bias and associated error, and lack of transparency.

The College also says close oversight is also vital because AI is being used for ever-more complex tasks, including planning for the treatment of tumors.

RANZCR President Professor John Slavotinek said: “The reality is that AI can revolutionise health care and when used properly, can free-up the workloads of specialists leading to more patients being seen and ending backlogs.

“We still have a responsibility as doctors to ensure AI medical tools remain safe in the clinical setting and are robustly monitored with continuous oversight.

“Humans and machines have quite separate ways of processing information - we must be aware that the brain works differently from the microchip.

“That’s why RANZCR is asking its members to closely supervise AI when using such tools in a clinical setting.

“As AI becomes more important in the medical world it is likely that further regulations about its use, such as those being introduced in Europe and the USA will become more prevalent.”