A Melbourne psychiatrist has reportedly refused to accept new patients who do not consent to having their consultations recorded and transcribed by an AI scribe.
The policy has sparked fresh concerns about patient privacy, informed consent and the lack of regulation governing AI use in Australian healthcare.
The case has been highlighted in a new report from Digital Rights Watch, which warns that AI scribes are rapidly becoming embedded across medical practices while laws designed to protect patients have failed to keep pace.
The report, “OFF THE RECORD: AI scribes in Australian healthcare”, argues that the growing adoption of AI transcription tools is creating a regulatory gap where patients may have little control over how highly sensitive health information is captured, processed and stored.
According to the Royal Australian College of General Practitioners, almost half of all Australian GPs now use AI scribes, with adoption doubling over the past year.
However, Digital Rights Watch claims many patients are being introduced to the technology without clear information about how their data is handled, whether they can refuse its use, or what safeguards exist if the technology makes mistakes.
The organisation reviewed 60 healthcare clinic privacy policies and found only four mentioned AI scribes.
It also contacted seven AI scribe providers about their data practices, with only two responding. Digital Rights Watch said none of the companies were willing to disclose the underlying AI models powering their systems.
The report raises concerns about transcription errors, algorithmic bias and the possibility of inaccurate information entering permanent medical records.
Digital Rights Watch said AI scribes have shown lower accuracy when processing patients with non-English accents, speech differences or minority backgrounds.
The organisation warned that AI-generated errors could potentially contribute to incorrect diagnoses, inappropriate medication decisions or inaccurate medical histories.
Tom Sulston, Head of Policy for Digital Rights Watch says we shouldn’t be forced to provide AI corporations with our most private information in order to get it. Everyone in Australia should be able to refuse the use of an AI scribe when they access healthcare.
“AI corporations operate in their own interests: not clinics’, not doctors’, and certainly not ours. Regulators need to work together to make AI scribes work for regular people.” said Sulston
“That means treating them as medical devices, increasing transparency and reporting requirements, and demonstrating that they work for all Australians equally.” he said.
Digital Rights Watch is calling for AI scribes used in healthcare to be regulated by the Therapeutic Goods Administration (TGA), with stronger guidance and enforcement from the Office of the Australian Information Commissioner (OAIC).
The organisation also wants privacy impact assessments to become mandatory before AI systems are introduced into clinical environments.
It is also calling for patients to have an explicit right to refuse AI scribes without affecting their access to healthcare.
Ed Santow, Director of the Human Technology Institute and former Human Rights Commissioner says the growth of AI scribes in health care has increased almost exponentially – with very limited regulatory oversight to protect patients and their personal information.
“There is a clear need for stronger oversight of this practice, and clearer rules to ensure that AI scribes may be used only with the free and informed consent of patients, and only in ways that uphold the community’s most basic health care and other rights.” said Santow
Meanwhile, a patient who shared their experience with Digital Rights Watch said they were not given the opportunity to refuse consent for an AI scribe during a healthcare appointment.
“When I shared my story with Digital Rights Watch I honestly didn’t think it was that big a deal, just yet another frustrating thing I came across while trying to access healthcare,”
“As it turns out this has generated a huge amount of discussion and it’s made me realise how serious it really was. I see it as part of a bigger picture – ‘forced inevitability,”
“The plan is clearly to slowly normalise this technology across society until it is impossible to opt out of, and it’s up to us to stop this happening.” the patient said.
Sophie Dutertre, Coordinator, Projects & Training at The Centre for Culture, Equity & Health, warned AI scribes could create additional barriers for vulnerable patients.
“We need to ensure that the use of AI scribes in health settings does not further disadvantage patients of culturally and linguistically diverse backgrounds, who speak with an accent or who use an interpreter. It is an equity issue.” she said.
The report comes as healthcare providers increasingly turn to AI tools to reduce administrative workloads, with advocates arguing the technology can give clinicians more time with patients.
However, Digital Rights Watch argues that efficiency gains must not come at the cost of informed consent, privacy protections and patient control over sensitive health information.
