AI in healthcare: How digital scribes could rewrite medical negligence law in Australia
30 October 2025
In summary
AI scribes are being used by nearly one in four Australian GPs to automatically transcribe patient consultations, creating new legal complexities around medical negligence liability when the technology makes errors or omits crucial information that doctors miss and correct.
Imagine you're at the doctor's surgery having a consultation. Your GP asks about your symptoms, you discuss treatment options, and throughout the conversation, artificial intelligence (AI) is quietly listening, processing, and transforming your words into structured medical notes – no human transcriptionist required.
This is the new reality of AI scribes, a technology that automatically transcribes and summarises patient consultations using advanced language models instead of traditional human notetakers. While doctors have long relied on dictation and human transcription services, AI scribes represent a fundamental shift: machines that can understand medical context, identify key information, and generate clinical documentation in real time.
Interestingly, nearly one in four GPs nationally are believed to be using these AI digital scribes for notetaking, according to the Australian Medical Association. But with convenience comes complexity – especially for medical negligence law.
What are AI scribes?
Think of AI scribes as sophisticated digital assistants. They
- listen to your consultation through a microphone,
- convert speech to text using speech recognition, and
- apply language models to extract the key medical details and generate structured notes.
Some can even draft referral letters and patient instructions automatically. The technology works in three steps: ambient listening during the consultation, transcription and processing of the conversation, and automatic generation of clinical documentation. The AI is trained to filter out small talk and focus on clinically relevant details.
The doctor's duty remains unchanged
A doctor's primary legal duty of care to their patient hasn’t changed. But AI scribes in healthcare add new responsibilities that could affect medical negligence claims:
Informed consent is essential. Recording a consultation without consent is a criminal offence in most of Australia. Before using an AI scribe, a doctor must get explicit, informed consent from the patient. This means explaining what the technology does, how the data will be used and stored, and who has access to it.
The buck stops with the doctor. Perhaps most relevant to potential medical negligence claims is a doctor’s ultimate responsibility for the accuracy of a patient's medical record. Medical records are the evidence on which medical negligence claims stand or fall. If the AI scribe makes a mistake, such as mishearing a key symptom, omitting a crucial detail, or misinterpreting a word due to a variation in accent, and the doctor fails to correct it, the doctor could be held liable. The law won't excuse a mistake just because a machine made it.
Clinical judgement still comes first. AI is a tool, not a substitute for medical expertise. Doctors must not become over-reliant on the technology. Doctors who rely on it and follow AI suggestions without applying their own expertise won’t meet the required standard of care.
A new world of evidential disputes
In many medical negligence cases against GPs, disputes often hinge on what was said during a consultation. Patients claim they mentioned a specific symptom, but the doctor's notes don't record it. AI scribes could completely change this dynamic with near-verbatim transcripts.
But it's a double-edged sword. On one hand, transcripts could prove that a doctor provided clear advice. On the other hand, they could strengthen patients' cases by showing crucial information was ignored.
The real challenge? When the AI gets it wrong. Legal disputes could shift from medical arguments to technical ones, requiring expensive AI expert testimony. For transcripts to hold weight in court, healthcare providers must prove recordings are authentic, time-stamped, and unaltered, demanding robust security and audit trails.
There's another catch: AI can't record what goes unsaid. A normally active child sitting motionless on their mother's lap, a patient's reluctance to discuss their home situation, these crucial non-verbal cues must still be manually documented.
Who's liable when things go wrong?
Traditionally, medical negligence claims focus on the doctor's actions. But now, AI introduces a new player: the software provider.
If an error in AI-generated notes leads to patient harm, is the doctor liable for missing the mistake, or is the software vendor liable for a faulty product? The answer depends on circumstances and isn't always clear.
Australia's ‘peer professional practice’ defence, established by the Civil Liability Act in each state or territory, allows doctors to show their actions were consistent with widely accepted practices. However, it becomes tricky when the technology is new and best practices are still emerging.
Privacy and regulatory uncertainty
AI scribes handle extremely sensitive health information, creating significant legal challenges.
Doctors and clinics must comply with Australia's Privacy Act 1988 and relevant state or territory health privacy laws, ensuring secure data storage and proper disclosure protocols.
The regulatory framework is still catching up. Most AI scribes aren't currently regulated as ‘medical devices’ by the Therapeutic Goods Administration (TGA), but this may change with time.
The rise of AI scribes promises to streamline healthcare, easing the paperwork burden for doctors and improving accuracy for patients. But the legal implications for medical negligence in Australia also demand a new level of scrutiny. For doctors, perhaps this means being more vigilant than ever, and for patients, it means being aware of their rights in a digitally enhanced consultation.
As AI becomes more common in consultations, the legal system will need to adapt fast and ensure patient safety remains the number one priority.
The ALA thanks Ric Traini for this contribution.
This is an edited version of an article first published by Maurice Blackburn Lawyers.
The views and opinions expressed in this article are the authors and do not necessarily represent the views and opinions of the Australian Lawyers Alliance.
Learn how you can get involved and contribute an article.