# AI in Kidney Education: Useful Work, Verifiable Evidence

AI can accelerate production; traceable sources and expert review establish trust.

![Infographic: AI in Kidney Education: Useful Work, Verifiable Evidence](https://urinenephrology.org/visual-reference/images/supplemental-ai-evidence.png?v=20261003c)

## Define the task

Use AI for a clearly specified educational or analytical job: organizing material, drafting explanations, extracting structured information, or generating teaching illustrations.

## Match the tool to the setting

Assess the intended use, data access, clinical consequences, and oversight needed. Predictive models and generative text systems have different capabilities and failure modes.

## Protect information

Use approved systems and follow applicable institutional privacy requirements. Removing obvious identifiers does not guarantee that a clinical narrative is fully de-identified.

## Verify every citation

Open the actual source and compare title, authors, identifiers, publication status, and relevant results. A convincing citation or a model-generated confidence score is not validation.

## Review the clinical reasoning

Check units, denominators, populations, dates, assumptions, and whether a guideline is final or draft. Separate trial findings, secondary analyses, and clinical extrapolation.

## Keep human accountability

Expert review, version control, and clear correction pathways support safe educational use. Measure whether the workflow improves accuracy and usefulness rather than assuming that speed proves benefit.

## Source lessons

- [AI healthcare](https://urinenephrology.org/AI_healthcare.html)

Read alongside the full lessons; the findings and decisions shown here require the stated clinical context.