Agentic education · Learning science

Actionable companion · Assessment & reasoning

Learning with agency: The Impact of Training Dental Students to Use an Artificial Intelligence–Based Platform for Pulp Exposure Prediction Prior to Deep Caries Excavation: A Proof‐of‐Concept Randomised Controlled Trial

AI-assisted editorial implementation guide, separate from the original research. Proposed actions are not claims 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 learning must be distinguished from K–12 instruction.

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

ABSTRACT Aim This study evaluated the effect of a short, personalised training session on student performance in using an artificial intelligence (AI)‐based platform for pulp exposure prediction before caries excavation and determined the required sample size for a further randomised controlled trial (RCT). Methodology Undergraduate dental students were randomly assigned to the experimental (training) group and the control (no training) group. The training group received a 1‐h training session before undertaking the experiment, focusing on the uses, applications, and drawbacks of AI and carious lesion penetration depth. The theoretical presentation was followed by practical exercises and a quiz to check learning progress. Later, participants in both groups completed an experimental task involving 292 cases. They were asked to predict pulp exposure using an AI‐based website. Sample size calculations determined the required sample size, with 80% power and an alpha of 5%. Results 18 participants were enrolled (9 in each group). The agreement between participants' decisions and AI predictions regarding the presence or absence of pulp exposure (agreeableness with AI) was higher in the training group compared to the control group (0.83 vs. 0.76). The training group had a slightly higher mean F1‐score (0.63 vs. 0.62), accuracy (0.69 vs. 0.68), and sensitivity (0.63 vs. 0.59) than the control group. Based on the sample size calculation, at least 31 participants per group are needed for the future RCT. Conclusions The results support further investigation of customised training sessions prior to using an AI‐based platform to assess their impact on dental students' agreement with AI predictions. Trial Registration ClinicalTrial.gov identifier: NCT05912361 Keywords: artificial intelligence, dental caries, randomised controlled trial

What kind of evidence is this?

Supporting AI learning science · Randomized study reported

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

The study followed the Preferred Reporting Items for Randomised Trials in Endodontics (PRIRATE) 2020 guidelines (Nagendrababu et al. 2020). This protocol was registered at ClinicalTrial.gov. Ethical approval was granted by the ethics committee of the related university (504‐0342/22‐5000), and data protection rules were followed (514‐0846/23‐3000) for using an AI‐based platform for predicting pulp exposure before deep caries excavation. The development of the AI model followed the checklist for AI in Dental Research (Schwendicke, Singh, et al. 2021).

Limitations and uncertainty · source passage

A previously developed DL model helped to predict the risk of pulp exposure before deep caries excavation using clinical data about age and pain and two‐dimensional radiographs (Ramezanzade, Dascalu, et al. 2023). While two‐dimensional radiographs have inherent limitations in detecting pulp exposure, they are still commonly used to estimate risk in clinical and research contexts (Gasqui et al. 2022). Although the model performed well, participants did not fully benefit from its output, likely due to a limited understanding of how AI generates predictions. To explore whether training improves interpretation and trust, a proof‐of‐concept randomised controlled trial (RCT) was conducted. Given the novelty of this topic, effect size data were lacking, justifying the need for this preliminary study (Gluud 2006; Haynes et al. 1997).

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

Shaqayeq Ramezanzade; Tudor Laurentiu Dascalu; Azam Bakhshandeh; Sergio E Uribe; Bulat Ibragimov; Lars Bjørndal. The Impact of Training Dental Students to Use an Artificial Intelligence–Based Platform for Pulp Exposure Prediction Prior to Deep Caries Excavation: A Proof‐of‐Concept Randomised Controlled Trial. International Endodontic Journal (2025-10-10). https://doi.org/10.1111/iej.70046

Companion prepared 2026-10-30. 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.