International Panel Sets Clear Rules for AI in Medical Writing
A consensus framework published in March 2026 establishes the first unified standard for how researchers can use artificial intelligence in medical papers. The guidance, developed by an international team led by experts from The First Affiliated Hospital of Shenzhen University, draws a clear distinction: AI can assist with language and structure, but cannot generate data, citations, or author claims.
The rules address a real problem. Most major publishers have AI policies, but they contradict each other. One journal demands every prompt and software version. Another simply says "be honest." This inconsistency leaves researchers uncertain whether using spell-checker AI crosses an ethical line.
The core prohibitions
AI cannot create or manipulate primary research images, including microscopy plots, gels, and flow cytometry data. AI cannot generate references. Large language models frequently fabricate authors, journal names, and DOIs-confidently presenting fiction as fact.
AI gets approval for supportive work: summarizing discussions, refining abstracts, improving readability in methods sections, and creating workflow diagrams. The condition is absolute: a human must verify every claim.
Section-by-section guidance
The framework includes a checklist tailored to each paper section. In Methods, AI can polish prose but cannot invent missing steps. In Results, text can be refined, but numbers must come from actual experiments. Non-data visuals like diagrams are permitted with full disclosure.
The authors tested their own rules. They used ChatGPT and Gemini during writing-only for language and structure, never for core science.
Transparency over prohibition
The authors rejected a policing approach. "We are not telling people to throw away their laptops," they said. The real danger is invisibility, not AI itself.
If a chatbot rewrites your discussion or suggests a citation, disclosure is required. The moment you claim AI-generated text as your own original thought, or a fabricated reference as real, you have committed misconduct. Transparency converts a black box into a tool.
The framework benefits non-native English speakers especially. They can now use AI for language help without ethical fear.
What this means for writers
Researchers gain a practical checklist to share with collaborators or include in supplementary files. Journals get a standardized audit trail for peer review. The guidance shifts the conversation from "is AI cheating?" to "how do we use AI well?"
For writers working in research and medical fields, AI for Writers resources can help navigate these emerging standards.
The authors committed to updating the rules as AI evolves. The consensus removes guesswork and protects science from silent errors while preserving the efficiency that modern writing tools provide.
Your membership also unlocks:
