What healthcare professionals should know first when using generative AI

Contents
  1. Jump to a topic
  2. Using generative AI in clinical practice
  3. Researching clinical questions and verifying references
  4. Using AI for grant applications, paper writing, and education

What should you check before entering patient information into generative AI? How do you verify the references AI cites? This page brings together articles covering practical examples and procedures—from clinical practice to grant applications, paper writing, and resident supervision. I add related materials as they become available.

Jump to a topic

This is the video from the seminar "Using generative AI to improve clinical quality." I discussed practical approaches to using generative AI in clinical practice and education. Details on how to get the presentation slides are in the lecture article below the video.

Watch the lecture on YouTube

[Video released] Using generative AI to improve clinical qualityshoei05.com

Using generative AI in clinical practice

What to check before entering patient information. Using draft referral letters, nursing care plans, and meeting minutes as examples, this separates what to check in opt-out training settings from what to verify in contracts and hospital policies.

Using generative AI in healthcare settings: Turning off model training is not always enoughshoei05.com

Drawing on reported cases, Dr. EARL outlines how to choose services and set hospital rules when drafting medical records with generative AI.

Precautions when using AI for medical charting: Staying compliant with guidelinesnote.com

An article covering domestic implementation examples and safety management issues for systems that transcribe clinical conversations and import them into electronic health records.

[Redefining Human-in-the-Loop in Medical AI] A thorough examination of direct voice-to-EHR AI integrationnote.com

Key issues when asking AI to generate differential diagnoses include handling patient information and how diagnostic accuracy varies with the input.

Pitfalls of AI differential diagnosis: Personal data handling and accuracy issuesnote.com

Using real outpatient examples, Dr. EARL examines how to respond when patients consult ChatGPT about their symptoms before coming to the clinic.

What if a patient comes to the clinic after asking ChatGPT about their symptoms?note.com

Researching clinical questions and verifying references

I built a Chrome extension to prioritize systematic reviews and meta-analyses when searching on OpenEvidence.

OpenEvidence Search Helper: systematic reviews, GRADE, and Japanese inputshoei05.com

UpToDate is an expert-edited resource, while OpenEvidence is an AI that searches the literature to generate answers. This summarizes how to choose between them based on the clinical question and why you need to verify AI answers against original papers.

The role of secondary sources (UpToDate) and search/general-purpose AI in answering clinical questionsshoei05.com

To make it easier to trace sources from Gemini's responses, I added prompt instructions to list cited references with links. The article includes the prompt.

How to include references in Geminishoei05.com

The Gem shared in Part 2 is configured to return a list of references with PubMed and DOI links when you enter a clinical question. The full instructions are included.

How to add references in Gemini (Part 2): Gem and prompt releasedshoei05.com


Using AI for grant applications, paper writing, and education

Before writing a grant application, you need to broaden your research question and read the literature to refine your hypothesis. This article introduces one AI agent to support each task.

Using generative AI to prepare grant applicationsshoei05.com

How do you verify references cited in a paper? What should you delegate to AI when developing research ideas? This article examines both questions, drawing on a guide to writing papers in English.

The ethics and potential of writing academic papers with generative AIshoei05.com

Reviewing together what a resident asked the AI and how they verified the basis for its answer. The article includes an example of a supervising physician and a resident working through this dialogue.


"Verify, then trust" — Fostering critical thinking in medical education in the AI erashoei05.com

Drawing on research into generative AI use in university courses and student critical thinking, I examined how to design assignments and assessments that require students to verify AI output.

Generative AI and critical thinking: Implications for university educationshoei05.com

When using AI-generated images in classes or teaching materials, it is important to check the copyright distinction between expression and ideas, as well as exceptions for educational use. This guide covers those points alongside ethical considerations when generating images.

Guide to using image-generating AI tools in educational institutionsshoei05.com


English translation of the Japanese article.