Commission on Excellence and Innovation in Health

Testing an AI scribe in South Australian emergency care

19 June 2026

Bearded male with glasses at a doctors appointment, the men are seated at the office desk

The CEIH, Cen­tral Ade­laide Local Health Net­work (CAL­HN) and the Aus­tralian Insti­tute for Machine Learn­ing (AIML) are work­ing togeth­er to tri­al a local­ly devel­oped AI-sup­port­ed clin­i­cal doc­u­men­ta­tion tool — often called an AI scribe — in CAL­HN emer­gency depart­ments and urgent care clinics.

The tool, known as AU Scribe, is designed to reduce man­u­al paper­work by auto­mat­i­cal­ly turn­ing patient-clin­i­cian con­ver­sa­tions into clin­i­cal notes. This allows clin­i­cians to spend more time focus­ing on patient care. 

The tri­al will assess whether a tool built on South Aus­tralian data, designed and test­ed with­in the SA Health sys­tem, can deliv­er val­ue. This means it can bet­ter reflect our local pop­u­la­tion, clin­i­cal work­flows and sys­tem require­ments. It also sup­ports local han­dling and gov­er­nance of sen­si­tive patient data, pro­vid­ing greater trans­paren­cy and con­trol over how infor­ma­tion is managed.

Now in its ear­ly Proof of Val­ue stage, AU Scribe is par­tial­ly fund­ed by the South Aus­tralian Gov­ern­ment Dig­i­tal Invest­ment Fund. The tri­al is test­ing whether patient-clin­i­cian con­ver­sa­tions in busy emer­gency depart­ments can be con­vert­ed (almost) auto­mat­i­cal­ly into mean­ing­ful infor­ma­tion and inte­grat­ed safe­ly and effec­tive­ly into clin­i­cal practice. 

Patient par­tic­i­pa­tion is vol­un­tary, with informed con­sent required. Strong safe­guards are in place. The tri­al includes over­sight of pri­va­cy, data man­age­ment and sys­tem per­for­mance, with input from clin­i­cal, tech­ni­cal and gov­er­nance stakeholders.

A key goal of the tri­al is to under­stand what’s need­ed for safe, scal­able imple­men­ta­tion in South Aus­tralia, includ­ing how the tool inte­grates with exist­ing sys­tems and meets pol­i­cy require­ments. Find­ings will inform future eval­u­a­tion and any poten­tial broad­er rollout.