Agentic education · Learning science

Actionable companion · Assessment & reasoning

From AI evidence to responsible practice: Using ChatGPT in Psychiatry to Design Script Concordance Tests in Undergraduate Medical Education: Mixed Methods Study

AI-assisted editorial implementation guide, separate from the original research. Proposed actions are not findings or recommendations attributed to the source authors. No independent peer review of this guide is asserted.

Read original paperSearch paper agentCC BY 4.0

Evidence and context

Title and abstract screening; stage labels describe reported settings, not verified numerical ages. Design labels require full-text appraisal. Reported stage(s): Higher education. Health-professions education findings do not establish effectiveness in K–12 classrooms.

Original authors’ abstract, reproduced under the linked source license:

Abstract Background Undergraduate medical studies represent a wide range of learning opportunities served in the form of various teaching-learning modalities for medical learners. A clinical scenario is frequently used as a modality, followed by multiple-choice and open-ended questions among other learning and teaching methods. As such, script concordance tests (SCTs) can be used to promote a higher level of clinical reasoning. Recent technological developments have made generative artificial intelligence (AI)–based systems such as ChatGPT (OpenAI) available to assist clinician-educators in creating instructional materials. Objective The main objective of this project is to explore how SCTs generated by ChatGPT compared to SCTs produced by clinical experts on 3 major elements: the scenario (stem), clinical questions, and expert opinion. Methods This mixed method study evaluated 3 ChatGPT-generated SCTs with 3 expert-created SCTs using a predefined framework. Clinician-educators as well as resident doctors in psychiatry involved in undergraduate medical education in Quebec, Canada, evaluated via a web-based survey the 6 SCTs on 3 criteria: the scenario, clinical questions, and expert opinion. They were also asked to describe the strengths and weaknesses of the SCTs. Results A total of 102 respondents assessed the SCTs. There were no significant distinctions between the 2 types of SCTs concerning the scenario (P=.84), clinical questions (P=.99), and expert opinion (P=.07), as interpretated by the respondents. Indeed, respondents struggled to differentiate between ChatGPT- and expert-generated SCTs. ChatGPT showcased promise in expediting SCT design, aligning well with Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition criteria, albeit with a tendency toward caricatured scenarios and simplistic content. Conclusions This study is the first to concentrate on the design of SCTs supported by AI in a period where medicine is changing swiftly and where technologies generated from AI are expanding much faster. This study suggests that ChatGPT can be a valuable tool in creating educational materials, and further validation is essential to ensure educational efficacy and accuracy.

What kind of evidence is this?

Supporting AI learning science · Survey / observational study

Stage: Higher education. Screening labels describe reported context and study-design language. They are not a risk-of-bias rating or proof of effectiveness for an age band.

Study methods · source passage

This study received the approval of the ethics of research committee of the Université de Montréal (approval 2023-4906). Participants were given a description of the study in the letter they received and were asked for their consent for their data to be used. Data were anonymized. The participants received no compensation for this study.

Limitations and uncertainty · source passage

In all, 2 (5%) of the 39 respondents mentioned that they found the SCTs straightforward and unchallenging. There were no other comments regarding the limitations of the SCTs generated by experts in the field. Sample responses included the following:

Design the system before the pilot

Use this study to inform a learning mechanism or implementation question. It is not direct evidence that autonomous agents improve learning. Keep the proposed agent behavior separate from what the source actually tested.

  • Use common rubrics, blind scoring where feasible and a fresh unaided reasoning task. Document permissible assistance and avoid treating AI-detector scores as proof of misconduct.

A 90-day district pilot

  1. Days 1–15 · Appraise. Have the curriculum lead and a teacher read the methods and limitations. Record the study design, sample, comparator, outcomes and evidence gaps. Define a local question that the study can help inform.
  2. Days 16–30 · Design. Choose one disciplinary task. Require a first attempt, a documented AI consultation, a critique of the output and an independent explanation. Align the comparison group and scoring rubric with the same objective. Name an accountable owner and set the baseline, success threshold and budget before implementation.
  3. Days 31–60 · Test. Start with a small voluntary cohort, preserve an accessible non-AI route and require teacher review. Use synthetic or de-identified examples; approve data handling and tool terms through district processes.
  4. Days 61–90 · Decide. Blind-score reasoning and accuracy on a new task without AI. Distinguish model examination scores, student perceptions and actual learner gains; only the last directly tests educational benefit. Compare with baseline, document uncertainty and stop or revise if learning worsens, access is unequal or critical errors persist.

The board decision brief

Report the learning problem, evidence quality, participating population, total cost including review time, measured benefit, errors, access differences and remaining uncertainties. Make a dated recommendation to stop, revise or expand. A positive perception survey alone does not establish learning improvement.

Attribution and reuse

Alexandre Hudon; Barnabé Kiepura; Myriam Pelletier; Véronique Phan. Using ChatGPT in Psychiatry to Design Script Concordance Tests in Undergraduate Medical Education: Mixed Methods Study. JMIR Medical Education (2024-04-04). https://doi.org/10.2196/54067

Companion prepared 2026-11-01. Original authorship and DOI are preserved. The abstract is an attributed source extract; the implementation plan is new AI-assisted editorial material, offered under the same CC BY 4.0 license. This guide has not undergone independent peer review.