Skip to main content

🎓 Practical Tips for Educators

Attribution

Original work: "Educators' guide to multimodal learning and Generative AI" — Tünde Varga-Atkins, Samuel Saunders, et al. (2024/25) — CC BY-NC 4.0
Adapted for UK Nursing Education by: Lincoln Gombedza, RN (LD)
Last Updated: July 2026

Here are five key strategies to help you integrate Generative AI effectively, ethically, and efficiently.

2026 Model Note

The examples below work with any current frontier model. As of mid-2026 the leading choices are Claude Sonnet 5 (Anthropic's default agentic model), Claude Opus 4.8 (deepest reasoning), GPT-5.5 (OpenAI), and Gemini 3.5 Flash/Pro (Google). These are reasoning models — they plan and think through problems on their own, so you no longer need to write "think step by step". Focus instead on a clear brief and good examples (see The Prompt Refinement Loop below).

🌟 Top 5 Educator Strategies​

1. Save Time ⏱️

Use AI for rote tasks like quiz generation.

See How

2. Mirror Use 🪞

Show students YOUR prompts and process.

See How

3. Checklists ✅

Always verify accuracy and ethics.

See How

4. Collaboration 🤝

Design for Human-AI partnership, not substitution.

See How

5. AI Literacy 📚

Teach prompt engineering as a core nursing skill.

See How

6. Prompt Loop 🔁

Refine prompts in a loop: Brief → Run → Critique → Verify.

See How

1. Save Time with smart use of AI Tools​

Streamline resource creation to focus on teaching.

📝 Quick Prompts​

Quiz Generation

Prompt

"Create 10 multiple-choice questions to assess Year 2 nursing students' understanding of hypertension management, aligned with NICE guidelines CG127. Include rationales for correct answers."

Care Pathway Flowchart

Prompt

"Convert this diabetes Type 2 management NICE guideline into a visual flowchart showing decision points for medication escalation. Use Mermaid diagram syntax."

2. Mirror AI Use: Show Your Process​

Model transparency. If you use AI to draft a lesson plan, tell your students.

"I show my students my ChatGPT conversation history... They see my prompts, the AI's responses, and my edits. It demystifies the process." — Nursing Educator

Activity: The "AI vs. Human" Comparison​

  1. Project: Show an AI-generated care plan.
  2. Compare: Show a nurse-written one.
  3. Discuss: What nuance did the AI miss?

3. Use Checklists to Guide Ethical & Sustainable AI Use​

Before hitting "Send" or "Print", run a quick audit.

✅ The "SAFE-AI" Checklist​

Clinical Accuracy
Ethical Red Lines

4. Design for Human-AI Collaboration, Not Substitution​

Think Co-pilot, not Chauffeur.

RoleAI Does...Student (Human) Does...
Care PlanningGenerates initial templatePersonalizes for patient needs
LearningCreates visual diagramsCritiques accuracy & relevance
InterviewsSimulates interviewerPractises verbal delivery

5. Teach AI Literacy as a Core Skill​

Don't assume digital natives are "AI natives".

Activity: The Prompt Challenge​

Goal: Show how specificity changes output.

  1. Bad Prompt: "Create a patient scenario."
  2. Good Prompt: "Create a realistic scenario for a 68yo patient with AF on warfarin. Include vital signs and INR results. Focus on bleeding risk."
  3. Compare: Have students analyse the difference in quality.

6. The Prompt Refinement Loop​

The single biggest shift since 2024: a good prompt is no longer a clever one-liner — it is a clear brief that you refine in a loop. Modern models (Claude Sonnet 5, GPT-5.5, Gemini 3.5) reward you for treating the conversation as an iterative cycle, exactly like the reflective Practice Loop you already use in clinical supervision.

Step 1 — Write a Brief, Not a Question​

Modern prompting reads like a job description, not a search query. State four things:

ElementExample (for a teaching scenario)
Role"You are an experienced UK adult-nursing lecturer."
Goal"Write a 400-word simulated patient handover for a Year 2 seminar."
Boundaries"Use only NICE/BNF-consistent information. No real patient data. UK spelling."
Format"Structure it as SBAR. End with 3 discussion questions."
Full Prompt

"You are an experienced UK adult-nursing lecturer. Write a 400-word simulated patient handover for a Year 2 seminar on sepsis recognition. Use only NICE NG51-consistent information, UK spelling, and fictional details (no real patient data). Structure it as SBAR and finish with 3 discussion questions that test the NEWS2 escalation pathway."

Step 2–4 — Show, Don't Just Tell​

The most reliable way to steer quality is a worked example ("few-shot prompting"). Paste one care plan you like and say "match this structure and tone." Then refine by reacting to the draft:

"Good, but the rationale for the fluid bolus is too vague — expand it and cite the relevant NICE guideline."

Each turn is one loop. Two or three loops almost always beats one long "perfect" prompt.

Step 5 — Verify Before You Trust​

The Loop Never Skips Verification

Every loop ends at the same gate: check every clinical claim against an authoritative source (NICE, BNF, Royal Marsden, NMC). Reasoning models are far more reliable than 2023-era chatbots, but they can still produce a plausible-sounding wrong dose. The nurse — not the model — remains accountable.

🚀 Action Plan: Getting Started​

Ready to dive in? Here is your roadmap.

📅 This Week

  1. Choose ONE teaching session to experiment with.
  2. Try generating a visual aid or quiz.
  3. Share the result with a colleague.

📅 This Month

  1. Develop a personal GenAI Checklist.
  2. Run one "Mirror Use" demo in class.
  3. Design one AI Literacy activity.

📅 This Semester

  1. Integrate AI Literacy into module outcomes.
  2. Build a shared Prompt Bank for students.
  3. Reflect on your own use in your PDR.

Next: See these tips in action with Nursing Examples!