Using AI for GP Trainee Portfolio Reflections: What's Allowed and What's Not

Featured image for Using AI for GP Trainee Portfolio Reflections: What's Allowed and What's Not

Over 60% of GP trainees already use ChatGPT for reflective practice. The RCGP knows. The question is not whether AI is used — it is how it is used.

What the RCGP Says

AI can support reflection — structuring your thinking, prompting deeper analysis, improving the clarity of your writing. AI must not replace reflection — the developmental purpose of reflective practice requires your own cognitive engagement. Overreliance on AI means you do not develop reflective skills — which are career-long professional skills, not just portfolio requirements.

What Is Acceptable

Using AI to generate reflection prompts based on your clinical encounter. Using AI to improve the clarity and structure of a reflection you have already written. Using AI to identify which RCGP capabilities your entry demonstrates. Using purpose-built tools (Learner+, iatroX CPD) that are designed for this workflow.

What Is Not Acceptable

Entering patient-identifiable information into consumer AI tools (ChatGPT, Gemini, Copilot). Submitting AI-generated reflections as your own without genuine cognitive engagement. Using AI for MSF self-reflection or other assessments where authenticity is essential.

The IG Non-Negotiable

Purpose-built clinical AI tools (Learner+, iatroX) have appropriate data handling for clinical contexts. Consumer AI tools do not. Never enter patient details — even anonymised ones — into ChatGPT. Your deanery's IG framework applies regardless of how the AI tool is marketed.

Transparency

Both RCGP and RCPCH recommend declaring AI use to your appraiser or supervisor. Transparency protects you. If a panel explores an entry and you have not disclosed AI assistance, the credibility impact is worse than if you had been upfront.

Panel Scrutiny

Expect some entries to be explored by your educational supervisor and/or ARCP panel. Formulaic language, identical structures, and generic learning points are red flags. If you cannot discuss the clinical content and learning from an entry in depth, the entry does not serve its purpose.

Where iatroX Fits

iatroX's CPD module provides AI-assisted reflection scaffolding within a platform designed for clinical data — avoiding the IG risks of consumer AI tools. The reflection content remains yours; the structure is scaffolded.

Share this insight