Students, early-mid career researchers, and clinicians have come together for a novel crowdsourcing event to identify risks associated with the use of artificial intelligence (AI) scribes in a health setting.
Organised by the Digital Health Cooperative Research Centre (DHCRC), the University of Sydney, and Lyrebird Health, over 20 attendees participated in the crowdsourcing workshop, observing the use of Lyrebird’s AI-enabled digital scribe in various clinical scenarios. The participants were then challenged to identify the risks they observed.
Furthering Safe, Ethical AI Use
The event was the first in a Digital Scribe Safety Challenge series that is designed to provide both a forum for furthering the safe and ethical use of Lyrebird’s AI Scribe in clinical settings, as well as providing an opportunity for students to build their knowledge and skills using emerging health technology, and assessing the effectiveness of a crowdsourcing model to identify risks associated with disruptive digital health technologies.
Annette Schmiede, CEO of DHCRC, said while take-up of AI scribes is growing among GPs and clinicians, there remains limited advice or guidelines on testing the safety of AI in healthcare.
It’s not enough to build something powerful – it has to be proven safe, trusted, and grounded in evidence
“There is limited practical advice or guidelines to test the safety of emerging digital health technologies, such as AI scribes,” Ms Schmiede said. “This represents a novel approach to solving this gap, providing practitioner feedback based on real-world scenarios.”
Lyrebird Health securely listens to patient consultations and then generates a tailored summary note, ready to upload to the patient’s medical record. It is used in over 60 hospitals in Australia, the United Kingdom, the United States, and Dubai. In Australia alone, it already documents over 150,000 consultations a week.
“Our mission has always been simple – to help doctors practice medicine, not paperwork,” said Lyrebird founder Kai Van Lieshout.
“This technology gives us an extraordinary opportunity to ease one of healthcare’s biggest pressures – the administrative burden that keeps clinicians from their patients. If we get it right, it can help make care more human again.”
“But developing technology like this comes with responsibility. It’s not enough to build something powerful – it has to be proven safe, trusted, and grounded in evidence. That’s why partnerships like this are so important: they bring clinicians, researchers and technology partners together to understand how these tools work in real practice, and how we can use them responsibly to truly support care.”
Valuable Student Development
University of Sydney Project Leader, Dr Anna Janssen, said the crowdsourcing workshops give students a unique hands-on opportunity to experience the role of AI tools in action.
“Using crowdsourcing is a novel way to support the ongoing improvement and safety of AI tools in the digital health landscape,” Dr Janssen said.
“This initiative gives students valuable, first-hand exposure to how AI scribes operate in real clinical settings, while strengthening their ability to identify and manage the risks associated with emerging technologies.”
An expert panel of researchers, health professionals, health service providers, health technologists, and other stakeholders interested in digital health will collate and prioritise risks identified in the crowdsourcing workshops and classify them against a harm framework. Up to three crowdsourced events will be run during the project.
