The French digital health startup Nabla has launched Copilot, a GPT-3 based tool designed to assist clinicians by transcribing patient conversations. Accessed initially as a Chrome extension, Copilot transcribes virtual conversations with patients and creates various endpoints such as prescriptions, follow-up appointment letters, and consultation summaries.
Why it’s notable
Nabla’s multidisciplinary team of doctors and machine learning engineers have been testing GPT-3 for various healthcare use cases since 2020. Nabla's Copilot represents a notable advancement, offering the potential to streamline administrative tasks, improve patient care, and reduce clinician burnout. Nabla's co-founder, AI entrepreneur Alexandre Lebrun, has a history of developing language-based applications, and was the co-founder of Wit.ai (acquired by Facebook) and VirtuOz (acquired by Nuance).
The tech is already being used by 20 digital and hybrid clinics across the U.S. and France and Nabla plans to launch an in-person consultation version of the tool in the coming weeks. Nabla has attracted over $20 million in funding, and while it was reported that it currently pays for GPT-3 as the basis of Copilot, it plans to develop its own large language model tailored to the unique language and needs of medicine and healthcare.
Industry Implications
The launch of Nabla’s Copilot comes at the right time for the healthcare sector as clinicians are often oversubscribed and face burnout. AI-powered tools like Copilot can help alleviate administrative burdens and minimise instances of human error. As healthcare digitisation has progressed, patients and clinicians are increasingly open to adopting digital tools to improve healthcare delivery. However, there are concerns about the risk of errors associated with large language models in healthcare, where even a small error rate can be unacceptable.
Nabla’s Copilot follows on from last months news from Doximity, a digital platform for medical professionals, who rolled out a beta version of a ChatGPT tool for doctors that helps streamline some of their time-consuming administrative tasks, such as drafting and faxing pre authorisation and appeal letters to insurers. Clinical documentation is a key opportunity for ChatGPT to improve workflows. The constraints of ChatGPT must be acknowledged. Even OpenAI, the creator of ChatGPT, admits that its output may contain inaccuracies.
Currently, ChatGPT's capabilities can be utilised to augment and enhance the skills and abilities of healthcare professionals and improve patient engagement, e.g. enabling them to get answers to their medical questions. As we move forward, it will be fascinating to see how the release of ChatGPT4 and other advanced AI technologies will shape the healthcare industry. The potential for ChatGPT to revolutionise patient care and improve health outcomes is vast. Overall, the healthcare industry is on the cusp of significant change, and ChatGPT and other advanced AI technologies are poised to play a crucial role in shaping its future. Relying too heavily on technology, it has been suggested that ChatGPT might diminish healthcare providers' capacity to diagnose and treat patients without relying on technology.