Artificial intelligence scribes—Are we ready?

Artificial intelligence (AI) has made the leap from blue-sky theory to bedside tool, yet many of us remain unsure how to use it safely. One application gaining traction is AI scribe technology—tools designed to generate medical notes from physician–patient conversations. The Doctors Technology Office recently published an article on choosing between the various machine learning models that can be tailored for clinical use.[1] Over the coming months, it will be interesting to see how our colleagues choose, or refuse, to integrate AI scribes and what lessons we might learn.

AI has been present in medicine for over 50 years. Early examples included programs like MYCIN, developed in the 1970s to diagnose and treat serious infections.[2,3] In more recent decades, AI applications in digital pathology[4] and radiology[5] have been proposed as physician partners, although implementation, ethics, and performance concerns remain.[6,7]

In April 2025, Doctors of BC released a policy statement on the use of AI in medical settings.[8] Among its commitments is a call for input from physicians, Indigenous Peoples, patients, medical learners, and other health care professionals to ensure all interest holders’ views are reflected. The statement also notes that although Health Canada uses a risk-based approach, “this framework has not been responsive to the evolving AI landscape. Common AI tools, like scribes . . . are largely unregulated.” It recognizes the need for a comprehensive set of standards—a laudable goal, but perhaps idealistic, given the pace of development in AI and the tremendous push for market adoption.

The College of Physicians and Surgeons of BC (CPSBC) published an interim guidance statement in October 2024.[9] It outlines six principles for using AI in practice:

1. Privacy, confidentiality, and consent, including familiarity with the Personal Information Protection Act as it relates to AI.
2. Accuracy and reliability—using critical thinking and clinical expertise when applying AI tools.
3. Transparency—being open with patients about the extent to which AI is used in their care.
4. Interpretability—understanding and interpreting AI-generated outputs appropriately.
5. Bias—recognizing issues around equity, diversity, and inclusion and algorithmic bias.
6. Monitoring and oversight—ensuring tools are used safely and appropriately over time.

The CPSBC makes it clear that these principles also apply to AI scribes, including the need for a physician to review documentation before entering it into the medical record. If the software stores audio recordings, physicians should reference the photographic, video and audio recording of patients practice standard.[10] The CPSBC also refers readers to additional guidance from the Law Society of British Columbia[11] and the Canadian Medical Protective Association.[12] After all, the responsibility for a medical record remains with the physician, even if AI drafts it.

While writing this editorial, I thought it appropriate to ask a large language model about integrating AI scribes into clinical practice. ChatGPT, perhaps surprisingly, began by suggesting that physicians start with a clear goal. Are you trying to improve patient flow? Reduce documentation burden and burnout? Setting an intention helps determine whether the intervention is actually useful. It also noted that AI cannot detect emotional tone or nonverbal cues—something we may take for granted, but worth emphasizing as we grow more reliant on AI-generated notes. Last, it recommended piloting the tool before a full rollout to gather feedback and make adjustments.

If you’re already using an AI scribe, the BCMJ would love to hear about it. How do you approach consent? Does AI enhance your workflow or communications? Has its presence changed how you supervise or teach learners? Have you encountered bias? How do your patients feel about being recorded?

In the title of this editorial, I didn’t ask “should we” use AI scribes, because their presence in medicine now seems inevitable. The real questions are: When will this happen? Who will shape how they’re used? Ultimately, it’s up to us—physicians—to define how these tools fit into our practices. And it may also fall to us to monitor, challenge, and guide AI developers to ensure these technologies evolve in ways that genuinely serve our patients.
—Caitlin Dunne, MD

hidden

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

References

1.    Doctors Technology Office. Using AI scribe technologies. Accessed 27 May 2025. www.doctorsofbc.ca/advice-support/doctors-technology-office/using-ai-scribe-technologies.

2.    Sotos JG. MYCIN and NEOMYCIN: Two approaches to generating explanations in rule-based expert systems. Aviat Space Environ Med 1990;61:950-954.

3.    Smith JW Jr, Svirbely J, Fannin E. A gentle introduction to knowledge-based systems in medicine. Top Health Rec Manage 1988;9:36-54.

4.    McGenity C, Clarke EL, Jennings C, et al. Artificial intelligence in digital pathology: A systematic review and meta-analysis of diagnostic test accuracy. NPJ Digit Med 2024;7:114. https://doi.org/10.1038/s41746-024-01106-8.

5.    Friedlander Serrano J. AI hasn’t killed radiology, but it is changing it. Washington Post. 5 April 2025. Accessed 27 May 2025. www.washingtonpost.com/health/2025/04/05/ai-machine-learning-radiology-software.

6.    Mello-Thoms C, Mello CAB. Clinical applications of artificial intelligence in radiology. Br J Radiol 2023;96:20221031. https://doi.org/10.1259/bjr.20221031.

7.    Parker W, Jaremko JL, Cicero M, et al. Canadian Association of Radiologists white paper on de-identification of medical imaging: Part 1, general principles. Can Assoc Radiol J 2021;72:13-24. https://doi.org/10.1177/0846537120967349.

8.    Doctors of BC. Artificial intelligence in health care. Policy statement. Updated April 2025. Accessed 27 May 2025. www.doctorsofbc.ca/sites/default/files/documents/2025-04-11-chep-ai-in-hc-policy-statement.pdf.

9.    College of Physicians and Surgeons of British Columbia. Ethical principles for artificial intelligence in medicine. Revised 3 October 2024. Accessed 27 May 2025. www.cpsbc.ca/files/pdf/IG-Artificial-Intelligence-in-Medicine.pdf.

10.    College of Physicians and Surgeons of British Columbia. Photographic, video, and audio recording of patients. Practice standard. Revised 7 October 2024. Accessed 27 May 2025. www.cpsbc.ca/files/pdf/PSG-Photographic-Video-Audio-Recording.pdf.

11.    Law Society of British Columbia. Guidance on professional responsibility and generative AI. Practice resource. Accessed 27 May 2025. www.lawsociety.bc.ca/Website/media/Shared/docs/practice/resources/Professional-responsibility-and-AI.pdf.

12.    Canadian Medical Protective Association. The emergence of AI in healthcare. Revised May 2023. Accessed 27 May 2025. www.cmpa-acpm.ca/en/advice-publications/browse-articles/2019/the-emergence-of-ai-in-healthcare.