Agentic education · Learning science

Actionable companion · Assessment & reasoning

From AI evidence to responsible practice: Physiotherapy students’ acceptance of AI-based chatbots (including ChatGPT) in education: a multi-institutional study from Turkey

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. Verify the participants, intervention and comparator before applying this study in a new setting.

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

Abstract Background AI-based chatbots are increasingly used in health professions education to support learning tasks (e.g., studying, drafting assignments, and summarizing content). Evidence on the acceptance of AI-based chatbots, including ChatGPT, in physiotherapy education is limited. This study examined physiotherapy students’ acceptance in Turkey, focusing on perceived usefulness (PU) and perceived ease of use (PEOU) within the Technology Acceptance Model (TAM). Methods A cross-sectional online survey (Google Forms) using convenience sampling was conducted with undergraduate physiotherapy students from nine universities across different regions of Turkey (May–June 2025). Digital informed consent was obtained, and responses were anonymous. PU and PEOU were assessed using Turkish TAM items contextualized for AI-based chatbot use. Psychometric properties were re-evaluated in the present sample (Cronbach’s alpha; PCA with KMO/Bartlett, with an EFA sensitivity analysis using PAF with Promax rotation), and group differences, correlations, and predictors of PU and PEOU were analyzed. Results A total of 478 students (79.3% female) were included. Mean PU and PEOU scores were 52.91 ± 7.47 and 51.73 ± 6.49, respectively (scale range 14–70). PU and PEOU were moderately correlated (r = 0.450, p < 0.001). Technology interest was the strongest predictor of PU (β = 0.283, p < 0.001) and PEOU (β = 0.371, p < 0.001), while year of study had a smaller effect on PU (β = 0.112, p = 0.011); gender was not significant in regression models. Exploratory factor-analytic results suggested PU was unidimensional, whereas PEOU showed two exploratory facets (learnability/clarity and control/flexibility). Conclusions Physiotherapy students expressed cautious optimism toward AI-based chatbots such as ChatGPT. Acceptance was higher among students with greater technology interest and, to a lesser extent, in advanced study years. These findings support stage-sensitive integration and targeted AI literacy, emphasizing ethical awareness and source verification for safe and effective learning. Supplementary Information The online version contains supplementary material available at 10.1186/s12909-025-08535-3. Keywords: Artificial intelligence, ChatGPT, Digital learning tools, Educational technology, Higher education, Physiotherapy education, Technology acceptance model

What kind of evidence is this?

Conversational & pedagogical agents · 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 cross-sectional quantitative study examined the perceptions of undergraduate physiotherapy students in Turkey regarding AI-based chatbots for academic purposes, within the TAM framework. This cross-sectional design was selected as a feasible approach for an early-phase, national snapshot of acceptance-related beliefs (PU/PEOU) across multiple institutions.

Limitations and uncertainty · source passage

This study has several limitations. The use of convenience sampling introduces the possibility of selection bias. Its cross-sectional design precludes causal inferences between variables. Self-reported measures may be subject to common-method bias. Because all variables were collected via a single self-report survey at one time point, common-method bias may have inflated some associations among constructs. We did not conduct a formal statistical test for common-method bias (e.g., Harman’s single-factor test). However, we attempted to reduce this risk procedurally through anonymous and voluntary participation and clear item wording. Future studies should apply recommended statistical and design-based remedies (e.g., marker variables, temporal separation, or multi-source measures).

Design the system before the pilot

Define whether the agent asks questions, gives hints, explains or assesses. Test hallucinated feedback, excessive praise, misleading confidence and responses when the learner is confused. A conversational interface alone does not establish autonomous agency.

  • 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

Mehmet Akif Güler; Sümeyra Dağ; Senem Kayal; Merve Can; Sena Nur Şimşekkaya. Physiotherapy students’ acceptance of AI-based chatbots (including ChatGPT) in education: a multi-institutional study from Turkey. BMC Medical Education (2026-01-03). https://doi.org/10.1186/s12909-025-08535-3

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.