Actionable companion · Motivation & engagement
From AI evidence to responsible practice: Sufficient and necessary conditions for ChatGPT adoption in medical education: a combined partial least square-structural equation modelling and necessary condition analysis
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.
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:
Background ChatGPT has gained rapid adoption in education due to its capacity to generate human-like responses, support personalized learning, and assist with complex knowledge retrieval. Yet, concerns about misinformation, overreliance, privacy, and ethical risks continue to shape students’ acceptance of AI tools. While prior research on factors influencing ChatGPT adoption has relied on regression-based approaches using frameworks like the Technology Acceptance Model, these methods assess only the sufficiency but not the necessity of these factors. This study combined Partial Least Squares Structural Equation Modelling (PLS-SEM) with Necessary Condition Analysis (NCA) to determine both the sufficient and necessary factors influencing medical students’ intentions to use ChatGPT for learning. Methods A cross-sectional survey was conducted among 146 pre-clinical Year 2 medical students at Universiti Malaysia Sarawak (UNIMAS). Seven variables were measured, i.e., perceived usefulness (PU), perceived ease of use (PEOU), social influence (SI), hedonic motivation (HM), perceived risk (PR), attitude (At), and behavioral intention (BI). Data were analyzed using SmartPLS 4.0, following a two-stage SEM approach, and NCA using the ceiling envelopment–full disposal hull (CE-FDH) technique. Results PLS-SEM showed substantial explanatory power (R² = 0.68) with attitude (β = 0.600), social influence (β = 0.133), and perceived usefulness (β = 0.096) as significant factors influencing intention for ChatGPT adoption. NCA revealed that attitude (d = 0.384), social influence (d = 0.310), and perceived usefulness (d = 0.183) were both necessary and sufficient conditions for BI whereas, hedonic motivation (d = 0.235) and perceived risk (d = 0.234), although non-significant in SEM, were significant necessary conditions, indicating threshold requirements of these 2 factors for adoption. Conclusion By combining PLS-SEM and NCA, this study enhanced the richness and granularity of understanding on the factors shaping ChatGPT adoption in medical education. Supplementary Information The online version contains supplementary material available at 10.1186/s12909-026-09244-1.
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 employed a quantitative, cross-sectional survey design among pre-clinical Year 2 medical students at Universiti Malaysia Sarawak (UNIMAS), Malaysia. Ethics approval was obtained from our institutional UNIMAS Human Research Ethics Committee (HREC) (with approval no HREC(NM)/2023 (2)/71) prior to the commencement of this study.
Limitations and uncertainty · source passage
This study has several pertinent limitations that should be acknowledged. First, as the sample was limited to a single cohort of pre-clinical Year 2 medical students from one institution, this might have limited the generalizability of findings to students in clinical years as well as students from other programs or educational settings. In this regard, a multi-center study across multiple year groups is therefore needed to confirm the generalizability of these results. Second, its cross-sectional design captures the students’ perceptions and behavioral intentions at a single time point only and would have prevented the capturing of causal relationship changes as the students’ experiences with ChatGPT evolves over a longer time period. Third, data were collected through self-administered questionnaires. This would have rendered the responses susceptible to social desirability and Hawthorne effects, where students might have overstated their positive attitudes toward ChatGPT. Fourth, the relatively homogenous Year 2 medical student cohort high baseline digital literacy and familiarity with these AI tools, could have reduced variability in perceived ease of use and contributed to the non-significant role of PEOU in this study. Finally, as the data were collected in December 2024, these findings reflect students’ perceptions of a specific ChatGPT version available at that time. Given the rapid iterations and frequent updates to the GPT model, newer versions may differ substantially in accuracy and functionality, thus potentially shifting perceptions and attitudes in future contexts.
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.
- Measure engagement separately from achievement. Check whether novelty-driven participation persists and whether it translates to independent performance.
- 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
- 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.
- 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.
- 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.
- 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
Keng Sheng Chew; Shirly Siew-Ling Wong; Say Keat Ooi; Zhi Jun Lim; Gregg Griffin Jensen; Nur Nayli Zulaikha Khairul Azri; Latchumy Shanmuganathan. Sufficient and necessary conditions for ChatGPT adoption in medical education: a combined partial least square-structural equation modelling and necessary condition analysis. BMC Medical Education (2026-04-29). https://doi.org/10.1186/s12909-026-09244-1
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.