[Video released] Using generative AI to improve clinical quality
In clinical settings, we often hear people say that we must not use generative AI. On social media, we also see examples of risky practices.
For example, some people enter full patient details—by voice or other means—into personal ChatGPT or Gemini accounts to generate summaries. Under current rules, this clearly violates guidelines and may also breach the Act on the Protection of Personal Information.
I am deeply concerned about these practices, which have no place in medical institutions and other professional settings.
At the same time, we are seeing more examples of generative AI in use. How should healthcare professionals approach and use generative AI on an individual level? I have released a video of a training session that addresses this question.
Moving toward institutional support
Under the FY2026 medical fee revision, using generative AI to help draft medical certificates and referral letters will be explicitly listed among the duties of medical administrative assistants. Specifically, using generative AI to help draft discharge summaries, medical certificates, and referral letters will qualify for evaluation under a mechanism that counts one medical administrative assistant as up to 1.3 staff members.
In other words, we are shifting from a stage where only interested individuals and institutions use AI to one backed by institutional systems.
Alongside generative AI, this revision covers voice input systems for medical documents, robotic process automation (RPA) for routine data entry, and explanatory videos for patients. It also requires compliance with the AI Guidelines for Business and the Guidelines for the Safety Management of Medical Information Systems, as well as training for assistants. Overall, the revision encourages adoption while requiring safe practices.
About the video
This video records a talk I gave on Friday, February 13, 2026, at Okayama University for the CoCoSimAI Pre-Seminar: "Latest updates on mastering generative AI to improve clinical quality."
It introduces practical ways to use generative AI in clinical practice and education, alongside the latest findings. I designed the talk for viewers with no prior knowledge of AI. I especially recommend it to anyone wondering how to put generative AI to practical use.
Download the handouts
You can download the presentation slides by completing the survey linked in the video description. Please feel free to use them in your daily clinical practice and teaching.
Conclusion
Medical fee revisions now explicitly recognize AI for streamlining administrative work. Building on this development, I hope to share practical knowledge through training sessions like this one.
Rather than simply saying "do not use it," it may be time to start thinking about how to use AI and trying it out. Institutional support is already underway.
I hope individuals will keep the necessary precautions in mind when using AI personally, and that organizations will follow the guidelines when integrating it into electronic health records.
I sincerely hope that appropriate and ethical use of generative AI spreads across medical institutions.