Artificial intelligence (AI) scribes can reduce the time physicians spend documenting patient visits by 69.1% during simulated primary care appointments, according to research led by University of Tor...

A study led by researchers at the University of Toronto found that artificial intelligence (AI) scribes substantially reduced the amount of time physicians spent documenting patient visits during simulated primary care appointments. The study, published in JAMIA Open, involved nine physicians who completed simulated encounters both with and without AI scribe technology. Researchers analyzed video recordings to compare how physicians allocated their time. When AI scribes were used, physicians spent 69.1% less time on clinical documentation. Without the technology, doctors devoted more than one-third of each simulated encounter to note-taking. AI scribes instead recorded the encounter and generated structured clinical notes that physicians could subsequently review and edit. Researchers observed that this workflow reduced interruptions during patient interactions and shifted documentation activity toward reviewing and correcting AI-generated notes. The researchers also identified an unexpected change in physicians' workflow. Some doctors used scratch pads to record thoughts and reminders during appointments, suggesting that conventional note-taking may support clinical reasoning and the formulation of follow-up questions. This finding highlights that reducing documentation time could also change how physicians organize information and make clinical decisions during consultations. The research is part of a broader University of Toronto and Women's College Hospital project examining AI technology in Canadian health care. The researchers emphasized the importance of studying the Ontario context because privacy regulations and health-care practices can differ from those in the United States. The findings do not establish how AI scribes will perform in routine clinical practice. The study used simulated encounters and involved only nine physicians. Further research is needed to examine real-world documentation quality, patient acceptance, privacy concerns, consent, and effects across different patient populations. Researchers also note that AI scribes address only one component of physicians' administrative workload, which can also include referrals, prescription orders, billing, and regulatory requirements.

Detectan datos inventados pero no lo que falta: la ceguera de omisión de los jueces LLM y los 2 métodos que la recuperan, según un paper de arXiv de 2026.

A growing number of Australian general practitioners are adopting AI scribes to streamline consultations, with approximately 40% using the technology. These tools transcribe patient interactions, generate notes, and draft care plans, reducing clinician workload and enabling more patient-centred care. However, experts highlight significant concerns, including potential errors in generated records, privacy risks, and regulatory gaps. While proponents argue AI scribes improve efficiency and reduce burnout, critics warn of 'hallucination'—where systems fabricate information—and the need for stricter oversight. Privacy advocates stress the importance of patient consent and data protection, with debates ongoing over whether AI scribes should be classified as medical devices. The article underscores the balance between innovation and accountability in healthcare technology.

Published in The American Journal of Bioethics (Vol. 26, No. 2, 2026)