Adelaide University’s Australian Institute for Machine Learning (AIML) has been awarded $1.9 million over three years from the Commonwealth’s Medical Research Future Fund (MRFF) National Critical Research Infrastructure scheme to expand the trial of AUScribe – dubbed as the first ambient AI medical scribe developed by an Australian government for its own public health system.
The project is led by Dr Anthony Porter and Associate Professor Emily Kirkpatrick as Chief Investigators, in a collaboration between AIML, the Central and Southern Adelaide Local Health Networks.
It seeks to build capacity across health, machine learning, ethics and implementation science, with the goal of supporting the growth of an Australian AI healthcare workforce in line with MRFF priorities.
The developers note that while Australian clinicians were “already adopting the technology at scale,” with up to 40 per cent of GPs using commercial AI scribes, those products process recordings in vendor clouds – interstate or offshore – and are designed for GP clinics and private practice.
AIML Chief Investigator Dr Anthony Porter said AUScribe differed from commercial products because as it stores all data on SA Health servers, “in full compliance with Australian public hospital data privacy law, and is purpose-built for public hospitals, including integration with electronic medical records.”
“It is also the only AI scribe trained exclusively on Australian public health data,” Dr Porter said.
“Commercial tools rely on US or European data, or partial Australian inputs. Training on SA Health data makes AUScribe more accurate and contextually appropriate for Australian clinical environments.”
He said the project includes bias testing for Aboriginal health representation, developed through Aboriginal-led co-design, to ensure the tool is safe and inclusive for Aboriginal and Torres Strait Islander patients.
Dr Porter noted that “no commercial scribe tested publishes comparable results,” adding that “unlike subscription-based commercial products, AUScribe is publicly owned, developed by AIML and the Central Adelaide Local Health Network with public investment.”
“For public hospitals that means long-term sustainability and affordability with no ongoing licence fees and no vendor lock-in – and an approach other states and territories can adopt.”
EVALUATION
The tool was currently being evaluated through a clinical trial in Australian emergency departments – including comparison with doctor-generated notes using established quality metrics.
“Most commercial scribes have limited peer-reviewed validation in the Australian hospital context. As a public sector initiative, the project operates under national AI governance guidance, with independent ethics approval and continuous oversight,” Dr Porter said.
The MRFF grant would support expansion across metropolitan emergency departments in the Central, Southern and Northern Adelaide Local Health Networks, with workflows designed to scale to inpatient wards, outpatient clinics and other specialties.
Dr Porter said because AUScribe was publically owned, “doctors and health networks can build their own AI applications on top of it, tailored to the way they actually work. You cannot do that with a commercial product. Doctors often feel disenfranchised from electronic medical records and health technology in general.”
“This is a way for them to lead the work and build it themselves. The partnership between CALHN and AIML makes this possible, combining frontline clinical experience with machine learning expertise. This is our opportunity to build public and professional trust in AI in healthcare, and we are doing it with the highest standards of care, transparency and rigour.”
SOVEREIGN
Associate Professor Emily Kirkpatrick, AUScribe Clinical Director and Chief Investigator said AIML had developed a sovereign, South Australian initiative, “that turns clinician-patient conversations into structured clinical notes and discharge letters, including a plain language letter for the patient and a letter for their GP.”
She added that it was built and hosted within the SA Health infrastructure and trained on local data, noting “This is a great working partnership with AIML, Adelaide University and the local health networks.”
AIML Chief Scientist Professor Anton van den Hengel said trial data illustrated that AI Scribe reduced documentation burden and allowed clinicians to spend more time with patients.
“Building our own AI means we can keep private patient data confidential, and this is critical, however, the real advantage to projects like this is that they provide an opportunity to create better outcomes for Australian patients,” Prof van den Hengel said.
“We are confident the ongoing trials will give us the data we need to build a robust and effective tool that ensures not only clinical accountability but also efficiency and efficacy.”
Add Comment