A faculty-supervised, ChatGPT-supported blended teaching model demonstrated higher short-term educational examination scores compared to conventional teaching during a four-week colorectal surgery internship, according to an exploratory cluster-randomized controlled study. The educational trial evaluated whether incorporating the generative artificial intelligence tool alongside traditional faculty supervision could improve learning outcomes for medical interns undertaking clinical training in colorectal surgery.
The clinical trial was conducted between June and September 2025 and involved a total of thirty interns organized into four pre-existing internship groups. Using an online draw procedure, investigators allocated two of the internship groups, encompassing fifteen students, to receive conventional surgical teaching, while the other two internship groups, also totaling fifteen students, were assigned to the faculty-supervised, ChatGPT-supported blended educational curriculum.
The research team defined their primary educational outcome as the overall course score, structured to reflect both theoretical understanding and clinical performance. This comprehensive outcome was composed of a final theoretical examination contributing fifty percent of the total mark and a practical examination contributing the remaining fifty percent, ensuring an equal balance between academic knowledge and applied procedural execution during the internship.
To capture practical proficiency, the practical examination was partitioned into three distinct evaluative components. A laparoscopic procedural assessment accounted for twenty percent of the practical score, while surgical suturing performance accounted for forty percent and clinical case analysis accounted for the final forty percent, establishing a structured assessment of technical execution and analytical decision-making in colorectal care.
The primary analysis compared all thirty participating students according to their allocated teaching arm. Interns in the conventional control arm recorded a mean overall course score of 76.77 with a standard deviation of 3.80 points. By comparison, students enrolled in the ChatGPT-supported blended teaching arm achieved a higher mean overall score of 86.85 with a standard deviation of 2.78 points.
At the individual student level, this performance gap translated into a statistically significant mean difference of 10.08 points, accompanied by a ninety-five percent confidence interval extending from 7.58 to 12.58 points and a P value of less than 0.001. In addition, the researchers assessed seven secondary educational outcomes, reporting Holm-adjusted P values that ranged from less than 0.001 to 0.003 across those measures.
Because the randomized allocation took place across pre-existing cohorts rather than completely independent individuals, the researchers also conducted a cluster-level sensitivity analysis. This cluster-level analysis identified a similar estimated mean difference of 10.13 points between the instructional arms; however, it generated a substantially wider ninety-five percent confidence interval ranging from −0.19 to 20.45 points, alongside an exact P value of 0.333.
In their concluding assessment, the authors observed that the ChatGPT-supported model was linked to superior short-term educational scores within the student-level evaluation. Nevertheless, they emphasized that because group allocation occurred across only four internship groups, the calculated statistical precision of the student-level analysis may be overstated, meaning that the overall findings must be interpreted strictly as exploratory rather than definitive.
Subject of Research: Faculty-supervised ChatGPT blended clinical teaching in colorectal surgery
Article Title: Application of generative artificial intelligence in clinical teaching of colorectal surgery: an exploratory cluster-randomized controlled study
Article References: Liu, Z., Wang, M., Xu, H., & Hu, S. (2026). Application of generative artificial intelligence in clinical teaching of colorectal surgery: an exploratory cluster-randomized controlled study. International Journal of Colorectal Disease. https://doi.org/10.1007/s00384-026-05231-6
Image Credits: AI Generated
DOI: 10.1007/s00384-026-05231-6
Keywords: ChatGPT, generative artificial intelligence, colorectal surgery, surgical education, cluster-randomized controlled trial, blended learning, surgical internship, laparoscopic assessment, suturing examination, case analysis, theoretical examination, practical examination, educational outcomes, cluster-level analysis, sensitivity analysis, clinical teaching, medical interns
News Source: Ophelia Keating. (November 19, 2026). ChatGPT Blended Learning Boosts Surgical Intern Scores in Exploratory Trial. Scienmag.



