Living guide · beta · last updated 1 November 2026 · Free for everyone. Missing something? Write to support@mrresearcher.com
Mr Researcher · Health AI Navigator
Lesson 8 of 12 · Applied

AI in research

About 25 minutes · plain language

In research, AI can play two very different parts: a helper in your work, or the thing you are studying. Each has its own guidelines. This lesson helps you decide which applies to your study and what you must report.

By the end you can
  • Choose the right reporting guideline for a study that uses or evaluates AI.
  • Write an AI-use disclosure for a manuscript.

Key terms

Reporting guideline
A checklist of what a study report should include so readers can judge and repeat it.Example: CONSORT for trials; CHART for chatbot studies.
AI extension
A version of a reporting guideline adapted for studies involving AI.Example: CONSORT-AI, SPIRIT-AI, STARD-AI.
Evidence synthesis
Combining the results of many studies, as in a systematic review.Example: A review of all trials of an AI tool.
Prompt record
The exact instructions given to a generative AI tool, kept so the work can be checked or repeated.Example: The full prompt and model version used to screen abstracts.
Authorship
Being named as an author, with responsibility for the work. AI cannot be an author.Example: Authors list AI use in the methods or acknowledgements.
AI helps you do researchSearching, screening, analysis, writing, translationGuidance: GAMER · ICMJE · RAISE (reviews)
AI is what you studyA model, chatbot, diagnostic tool or AI interventionGuidance: TRIPOD+AI · CHART · STARD-AI · CONSORT-AI · DECIDE-AI
First decide which part AI plays in your study; the guideline follows.

First decide: helper or subject?

AI as a helper means you use AI to do part of your research: searching, screening, analysing, translating or writing. AI as the subject means your study is about an AI tool: you build it, test it, or evaluate it in care. Many studies are both. The figure shows which guidelines apply to each.

When AI is your helper

You remain accountable for every result. Keep a record of each AI-assisted step, check its output, and disclose substantial use in the methods or acknowledgements, following the journal’s policy and the GAMER guideline. ICMJE and COPE agree that AI cannot be an author.

AI in systematic reviews

RAISE 2026 is living draft guidance from Cochrane, Campbell and partners on using AI in evidence synthesis. It sets out which review tasks AI may support and how to validate them. It currently rates large language models synthesising findings across studies as not acceptable, and every AI-assisted step needs human oversight.

When AI is the subject

Choose the reporting guideline that matches your design:

  • Prediction model (risk scores, machine learning): TRIPOD+AI.
  • Diagnostic accuracy of an AI test: STARD-AI.
  • Chatbot giving health advice: CHART.
  • Wider studies of large language models: TRIPOD-LLM.
  • Trial of an AI intervention: SPIRIT-AI for the protocol, CONSORT-AI for the results.
  • First live clinical use of AI decision support: DECIDE-AI.
  • Medical imaging AI: CLAIM.
  • Health economic evaluation of an AI tool: CHEERS-AI.

Why the AI version matters

The general guidelines (CONSORT, STARD, TRIPOD) do not ask for things that matter for AI, such as the algorithm version, how input data were handled, how people interacted with the output, and how errors were analysed. Reviewers increasingly expect the AI extension.

Case: is the chatbot’s breastfeeding advice safe?

A midwifery team wants to test how well a popular chatbot answers mothers’ questions about breastfeeding. Following CHART, they record the model name and version, the exact date, and every prompt word for word; they decide in advance how two midwives will score accuracy and safety; and they report every unsafe answer. In the methods they also disclose that AI helped translate the questions into Arabic.

Common mistakes

  • Using AI in a review without saying so.
  • Choosing the general guideline and missing its AI extension.
  • Reporting a chatbot study without the model version and prompts.

Check your understanding

1. Your study builds a machine-learning model to predict falls. Which guideline do you follow?

TRIPOD+AI.

2. You used ChatGPT to improve the English of your paper. What should you do?

Check the edited text, keep a record, and disclose the use according to the journal’s policy (GAMER describes what to report).

3. Why is CONSORT alone not enough for a trial of an AI tool?

It does not ask for AI-specific details such as the algorithm version, input data handling and human-AI interaction; CONSORT-AI does.

Try it

  • For your current or next study, decide whether AI is a helper, the subject, or both, and name the guideline you will follow.

Remember

  • First decide: is AI your helper or your subject?
  • Helper: record, check, disclose (GAMER, ICMJE).
  • Subject: use the AI extension that matches your design.
References
  1. Luo X, Tham YC, Giuffre M, et al. Reporting guideline for the use of Generative Artificial intelligence tools in MEdical Research: the GAMER Statement. BMJ Evid Based Med. 2025;30(6):390-400. doi.org
  2. International Committee of Medical Journal Editors. Recommendations for the conduct, reporting, editing, and publication of scholarly work in medical journals. Updated January 2026 (Section V: Use of artificial intelligence). www.icmje.org
  3. Committee on Publication Ethics. Authorship and AI tools: COPE position statement. COPE; 2023. publicationethics.org
  4. Thomas J, Flemyng E, Noel-Storr A, et al. RAISE: Responsible use of AI in evidence SynthEsis (RAISE 1–3), version 4, draft for consultation. OSF; 2026. osf.io
  5. Huo B, Collins GS, Chartash D, et al. Reporting guideline for chatbot health advice studies: the CHART statement. JAMA Netw Open. 2025;8(8):e2530220. doi.org
  6. Collins GS, Moons KGM, Dhiman P, et al. TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods. BMJ. 2024;385:e078378. doi.org
  7. Sounderajah V, Guni A, Liu X, et al. The STARD-AI reporting guideline for diagnostic accuracy studies using artificial intelligence. Nat Med. 2025;31(10):3283-9. doi.org
  8. Gallifant J, Afshar M, Ameen S, et al. The TRIPOD-LLM reporting guideline for studies using large language models. Nat Med. 2025;31(1):60-9. doi.org
  9. Liu X, Cruz Rivera S, Moher D, et al. Reporting guidelines for clinical trial reports for interventions involving artificial intelligence: the CONSORT-AI extension. Nat Med. 2020;26(9):1364-74. doi.org
  10. Cruz Rivera S, Liu X, Chan AW, et al. Guidelines for clinical trial protocols for interventions involving artificial intelligence: the SPIRIT-AI extension. Nat Med. 2020;26(9):1351-63. doi.org
  11. Vasey B, Nagendran M, Campbell B, et al. Reporting guideline for the early-stage clinical evaluation of decision support systems driven by artificial intelligence: DECIDE-AI. Nat Med. 2022;28(5):924-33. doi.org
  12. Tejani AS, Klontzas ME, Gatti AA, et al. Checklist for Artificial Intelligence in Medical Imaging (CLAIM): 2024 update. Radiol Artif Intell. 2024;6(4):e240300. doi.org
  13. Elvidge J, Hawksworth C, Avsar TS, et al. Consolidated Health Economic Evaluation Reporting Standards for interventions that use artificial intelligence (CHEERS-AI). Value Health. 2024;27(9):1196-205. doi.org

Lesson text is our own plain-language summary of these sources; it is not endorsed by their authors. Check the original before relying on any detail.