Wrong Answers Research Articles

The Collision between Tradition and Technology: The Impact and Enlightenment of AI on Chinese Education in China

At present, the application of AI in the field of education has been widely discussed, but there are still shortcomings in the combination of Chinese education and AI. This paper analyzes the influence of AI on Chinese education. And found that AI can help Chinese teachers reduce repetitive work, reduce the burden of preparing lessons, and understand students’ writing levels. However, improper use of AI will also lead to the weakening of teachers’ functions and the abuse of students’ integrity. For students, AI is not only an auxiliary tool to help students but also provides a learning platform. But it also causes problems such as solidified thinking, widening the gap between college students, and difficulty identifying wrong answers. Based on this, this paper puts forward the following suggestions. Chinese teachers should actively follow the trend and correctly use AI to improve classroom efficiency. Students should use AI reasonably, and they should not be dependent on it. Schools take incentive measures. The government should set up relevant policies in favor of AI in Chinese education.

Investigating Translation Students’ Challenges in Rendering Arabic Idioms and Proverbs into English: Insights and Recommendations

This study investigates the challenges faced by 4th-year students majoring in Translation at Al-Buraimi University College in Oman when translating proverbs and idioms from Arabic into English. These expressions pose particular difficulties in translation due to their cultural and linguistic specificity, often requiring more than literal translation to convey their intended meaning accurately in the target language. Understanding these challenges is crucial for enhancing the effectiveness of translation education programs and improving students’ proficiency in handling idiomatic expressions. The findings showed that 4th-year translation students experienced three main difficulties in translating proverbs and idioms, ranging from the semantic and syntactic fields to cultural differences. First, semantically speaking, they struggled to grasp the meaning of the source text and find the closest equivalent proverb or idiom in the English language. Second, from the syntactic perspective, learners showed several deficiencies in making a correct structure in English. Third, the high percentage of wrong answers revealed a lack of cultural awareness regarding the target culture. This research employs qualitative methods to identify common difficulties that students encounter, offering insights into potential strategies and pedagogical approaches to address these challenges effectively. By focusing on idiomatic expressions and proverbs, this study contributes to the broader discourse on translation education and highlights the importance of cultural competence in achieving accurate and nuanced translations.

Web-based training and certification of clinical staff during the randomised clinical trial SafeBoosC-III

BackgroundSafeBoosC-III is a pragmatic, multinational clinical trial evaluating cerebral oximetry-guided treatment for extremely preterm infants. In total, 1601 infants were randomised across 70 sites in Asia, Europe, and USA. To enhance data quality and patient care, a web-based training program was implemented for staff. We now report on the processes.MethodsAll training modules consisted of initial learning material followed by a case-based quiz, with elaborate responses to correct as well as to wrong answers. Modules covered trial introduction, cerebral oximetry monitoring, treatment guidelines, cerebral ultrasound, and Good Clinical Practice. The introduction module was accessible in eight languages on an online platform, while language versions varied for other modules, due to different needs. Certification was earned upon module completion, relevant to the staff category. The training was not mandatory, but for motivational purposes, principal investigators continuously received local certification rate reports.ResultsA total of 926 out of 2347 staff (39%) obtained certification. Amongst 295 staff who completed the evaluation, 83% rated the program as overall good and 94% found it relevant to clinical practice. Sites exhibited varying certification rates, with 10 at 0%, 43 between 0.1 and 79.9%, and 17 exceeding 80%. There was no correlation between the rate of certification in individual sites and how often the clinical management was changed due to cerebral hypoxia nor a correlation to site-specific estimates of the intervention effect.ConclusionDespite language barriers and a low budget, our web-based training and certification program proved feasible. Only a minority of sites reached 80% certification of staff and an impact on the trial could not be detected.Trial registrationThe SafeBoosC-III trial is registered at ClinicalTrials.gov NCT03770741. The first participant was randomised in June 2019 and recruitment was completed in December 2021.

Using Illustrated Children’s Book to Strength Fracted / Cracks Identity of (the Israeli) Children

Illustrations in children's books serve as powerful aesthetic devices to address complex emotions and ideologies, particularly in war-torn regions. These visual narratives act as catalysts for moral, value-based, and political discussions, offering a wordless medium that invites diverse interpretations based on the observer's perspective. This approach creates an open dialogue where there are no absolute right or wrong answers. Children's literature, through its illustrations, becomes a significant agent of socialization. These books impart cultural values, ideological concepts, and political behavior patterns, functioning as tools for establishing culturally responsive pedagogy. The images discussed, while originating from Israeli children's books predating the 2023 conflict, possess a timeless and universal quality that transcends their immediate context. Illustrations in children's books can either complement or diverge from the accompanying text, offering varying degrees of precision or poetic license. They carry cultural connotations that encompass symbols, traditions, beliefs, behaviors, and values unique to a particular society, conveying information beyond the apparent narrative. The inherent power of illustration lies in its ability to communicate directly with young children, often without the need for adult mediation. In preschool literature, illustrations serve multiple functions: they parallel written words, demonstrate abstract concepts, provide concrete representations of characters and settings, emphasize or contradict textual elements, and expand the narrative through visual elements not present in the text. Moreover, illustrations possess a psychological-therapeutic function, contributing to indirect treatment techniques. Like all visual representations, they embody simplified messages and shape recognition, conveyed formally and tangibly, with each element having both conscious and unconscious effects on the Therapeutic visual qualities in illustrated books aim to capture emotional nuances, intensify expressions of feeling, and convey concepts that may be challenging to express verbally. This is achieved through original solutions such as symbolism and creative visual metaphors. In conclusion, illustrations in children's books, particularly those addressing complex themes like war and fear, offer a unique and powerful medium for fostering understanding, encouraging dialogue, and providing emotional support. Their ability to transcend linguistic and cultural barriers makes them invaluable tools in shaping young minds and addressing challenging societal issues.

Flight Arrival Scheduling via Large Language Model

The flight arrival scheduling problem is one of the critical tasks in air traffic operations, aiming to ensure that the flight arrive in the correct sequence safely. Existing methods primarily focus on the terminal area and often overlook the presence of training flight at the airport. Due to the limited generalization of traditional methods and varying control practices at different airports, training flight at airports still rely on manual control for arrival sorting. To effectively address these issues, we propose a novel method for slot allocation that leverages the strong reasoning capabilities and generalization potential of large language models (LLMs). Our method conceptualizes the dynamic scheduling problem for training flight as a language modeling problem, a perspective not previously explored. Specifically, we represent the allocator’s inputs and outputs as language tokens, utilizing LLMs to generate conflict-free results based on a language description of requested landing information and assigned training flight information. Additionally, we employ a reset strategy to create a small dataset for scenario-specific samples, enabling LLMs to quickly learn allocation schemes from the dataset. We demonstrated the capability of LLMs in addressing time conflicts by evaluating metrics such as answer accuracy, conflict rate, and total delay time (without the wrong answer). These findings underscore the feasibility of employing LLMs in the field of air traffic control.

User Experience of ChatGPT: Insights from Undergraduate English Language and Literature Students in Bangladesh

ChatGPT, a large language model, has become increasingly popular, and students have begun to use it for the purpose of learning. The purpose of this research is to explore the experiences of undergraduate English language and literature students from Bangladesh who are using ChatGPT. Qualitative research methods were utilized for this study and the data were collected using semi-structured interviews. A thematic analysis technique was employed to analyze the data. The study reveals both positive and negative experiences associated with the use of ChatGPT by students from the English Language and Literature department. ChatGPT helps students to practice their reading and writing skills. It aids them to provide literary explanations of their texts and helps students to understand complex literary texts in lucid language. However, it sometimes provides confusing or wrong answers and provides wrong explanations regarding literary texts. As students possess some negative experiences along with many positive experiences, they are inclined to continue using it with heightened awareness of their use. However, users in this study believe that its use should be limited, and it should be used as a complement rather than a primary tool. The findings of this research may provide educators and policymakers with valuable insights that might help them integrate new aspects of artificial intelligence in education.

EXAMINING NUMBER SENSE SKILLS OF KINDERGARTEN CHILDREN

The aim of this study is to determine the number sense skills of preschool (Kindergarten) children. In the study, survey design, which is one of the quantitative research methods, was used. The study population of the research consists of a total of 114 children attending all of kindergartens (5 kindergartens) in the city center of Tunceli in the 2020-2021 academic year. The data collection tool used in the study is the "Preschool Number Sense Test" developed by the researchers. In the analysis of the data, the items in the number sense test; The correct answer obtained by using number sense skills, the correct answer obtained by using rule and operation-based methods, the wrong answer obtained by using number sense skills, the wrong answer obtained by using rule-operation-based solutions, and the blank answer were coded. As a result of the analysis made, it was determined that preschool students used rule- and operation-based solutions more in solving questions, and it was seen that they resorted to number-sense-based solutions less frequently.

Larger and more instructable language models become less reliable.

The prevailing methods to make large language models more powerful and amenable have been based on continuous scaling up (that is, increasing their size, data volume and computational resources1) and bespoke shaping up (including post-filtering2,3, fine tuning or use of human feedback4,5). However, larger and more instructable large language models may have become less reliable. By studying the relationship between difficulty concordance, task avoidance and prompting stability of several language model families, here we show that easy instances for human participants are also easy for the models, but scaled-up, shaped-up models do not secure areas of low difficulty in which either the model does not err or human supervision can spot the errors. We also find that early models often avoid user questions but scaled-up, shaped-up models tend to give an apparently sensible yet wrong answer much more often, including errors on difficult questions that human supervisors frequently overlook. Moreover, we observe that stability to different natural phrasings of the same question is improved by scaling-up and shaping-up interventions, but pockets of variability persist across difficulty levels. These findings highlight the need for a fundamental shift in the design and development of general-purpose artificial intelligence, particularly in high-stakes areas for which a predictable distribution of errors is paramount.

Adversarial Sample Synthesis for Visual Question Answering

Language prior is a major block to improving the generalization of visual question answering (VQA) models. Recent work has revealed that synthesizing extra training samples to balance training sets is a promising way to alleviate language priors. However, most existing methods synthesize extra samples in a manner independent of training processes, which neglect the fact that the language priors memorized by VQA models are changing during training, resulting in insufficient synthesized samples. In this paper, we propose an adversarial sample synthesis method, which synthesizes different adversarial samples by adversarial masking at different training epochs to cope with the changing memorized language priors. The basic idea behind our method is to use adversarial masking to synthesize adversarial samples that will cause the model to make wrong answers. To this end, we design a generative module to carry out adversarial masking by attacking the VQA model, and introduce a bias-oriented objective to supervise the training of the generative module. We couple the sample synthesis with the training process of the VQA model, which ensures that the synthesized samples at different training epochs are beneficial to the VQA model. We incorporated the proposed method into three VQA models including UpDn, LMH and LXMERT, and conducted experiments on three datasets including VQA-CP v1, VQA-CP v2 and VQA v2. Experimental results demonstrate that a large improvement of our method, such as 16.22% gains on LXMERT in the overall accuracy of VQA-CP v2.

Analysis of Question Items Using the Differentiating Power Method

The use of multiple choice questions in exams remains a common choice in education for various reasons, such as ease of assessment, perceived objectivity, and the ability to provide rapid feedback especially in large classes. Research shows that the use of multiple choice questions can strengthen retention of information, especially involving alternative wrong answers, as well as measure students' basic understanding in various subjects. However, to create quality multiple choice questions, an in-depth evaluation of the question elements is required, including item analysis to ensure the validity, reliability and fairness of the assessment. The results of this research using quantitative descriptive methods show that most of the questions can be improved, while a small number need to be rejected. The research conclusions suggest that rejected items should not be reused, while items that need to be corrected should be improved to improve the overall quality of the exam. Thus, analyzing the quality of multiple choice questions is crucial for increasing the effectiveness of assessment, especially in higher education contexts such as nursing and medical education.

Tingkat Pengetahuan Tentang Vitamin C Di Apotek Al Fayadh Farma 1 Bantul Yogyakarta

Vitamins are substances needed by the body because they have important functions for the body, vitamin C also has a role in the body's metabolic activities. Vitamin C deficiency can cause deficiencies in various diseases, such as mouth ulcers, anemia, dry skin and gingivitis. The aim of this research was to determine the level of knowledge regarding the use of vitamin C at the Al-Fayadh Potorono Pharmacy, Bantul. This research uses descriptive observational methods and uses prospective research methods. The sample used is 110 respondents. Data collection used a closed questionnaire. Data analysis used the Guttman scale where correct answers were given a score of 1 and wrong answers were given a score of 0, and categorized by percentage into 3 categories, namely good 76-100%, sufficient 56-75%, and less <56%. Based on the research results, it shows that the level of patient knowledge regarding the use of vitamin C is in the good category, 67 respondents (61%), sufficient category is 38 respondents (34%),and less than 5 respondents (5%). So it can be concluded that the level of knowledge of patients at the Al Fayadh Farma Pharmacy regarding the use of Vitamin C is good

Teaching with an adaptive technology

In this paper we present the first findings from a 4 year longitudinal study that investigates the interplay between an adaptive learning technology and educators and students in a nurse education in Denmark. With inspiration from grounded theory we analyse data from 14 hours of classroom observations supplemented by educator interviews, student surveys and mini interviews with students. In the paper we focus on the role of a new technology in the classroom: we ask how data reports from an adaptive learning technology influence educators’ preparations for classroom activities, and we ask how students' technology acceptance evolves over time. Analyzing our data we see three main themes emerge that contributes to answer the research question: Interpretative inclinations: The adaptive technology contributes to planning the lessons with a recurring inclination in practice: wrong answers lead to more educator presentations, and correct answers lead to student-centred group work during class. Implicit comparisons: The use of data reveals the educator’s expectations and assumptions when she implicitly compares the data on the students with her intuition, and when she compares the technology with the physical book. The implicit comparisons make the use of data challenging for the educator. Reverse adoption: The students find it hard to grasp the intention of the adaptive technology, but they accept and learn to adapt and appreciate it over time. In contrast to e.g. Rogers’ technology acceptance theory we find that the adaptive learning technology cannot be evaluated per se: in order to describe how a learning technology is accepted and what role it comes to play in an educational context, the parameter of time is mandatory to include in the research design.

The Roles of Pedagogical Agent’s Emotional Support: Dynamics Between Emotions and Learning Strategies in Multimedia Learning

The present research conducted two experiments with an intelligent tutoring system to investigate the overall and dynamic impact of emotional support from a pedagogical agent (PA). In Experiment 1, a single factor intergroup design was used to explore the impact of PA’s emotional support (supportive vs. non-supportive) on learners’ emotions, intrinsic motivation, and learning gain. Sixty participants were recruited and randomly assigned to one of the two conditions. Experiment 2 also conducted a single factor between-subjects design to investigated the dynamic patterns between emotions and learning strategies among 30 participants using lag sequential analysis. Results showed that: Compared with the non-supportive pedagogical agent, the supportive pedagogical agent reduced frustration and improved learning gain, but did not increase intrinsic motivation. In addition, learners with the supportive pedagogical agent used more appropriate strategies after frustration and surprise, and use less ineffective strategies after confusion and enjoyment to avoid reaching a wrong answer. If learners did not receive emotional support in such cases, learning strategies following these emotions were more likely lead to negative cognitive results, or negative emotions tended to appear repeatedly. Instructors or PAs should identify learners’ emotions in time and provide the appropriate emotional support according to learners’ emotions.

Generative AI Chatbot for Engineering Scientific Journal

Abstract: This study focuses on exploring the potential of artificial intelligence as an alternative, effective, and user-preferred tool for answering inquiries, compared to traditional methods such as FAQs or email and ticketing systems. The study highlights how AI can enhance efficiency and accuracy in processing and responding to inquiries from readers, authors, and reviewers, by providing immediate and customized answers based on the analysis of information available on the journal's website and the data fed to the chatbot. Through in-depth discussions and an analysis of the inquiries received over a full six months, totaling about 3000 inquiries, the study demonstrates the good ability of the chatbot to understand complex inquiries and provide satisfactory answers. The study indicates that chatbots can reduce the workload on editorial teams of scientific journals by automating responses to routine inquiries, allowing staff to dedicate more time to editorial and academic tasks. One of the key aspects of training is teaching the chatbot to provide correct answers to various inquiries and to avoid responding to negative or redundant inquiries. The research explores the challenges of applying AI in this context, including the need to train smart models to understand specific academic language and ensure accuracy in responses, as well as addressing privacy concerns and data security. The importance of designing flexible and adaptable AI systems to meet the diverse requirements of different scientific journals and their users is emphasized. The study concludes that artificial intelligence is a promising tool for improving the interaction between academic journals and their communities, offering an effective alternative to traditional systems. It highlights the necessity for ongoing research and development to enhance AI capabilities. Notably, the AI tool currently lacks a direct method for correcting its wrong answers, which is one of the most effective learning tools used by parents to correct their children's answers. One of the key recommendations of the study is that AI training should be conducted in stages.

Interviewing students with special needs: Developing ethical considerations and interviewing protocols

AbstractInterviewing students with special needs (SSNs) comes with a unique set of ethical and practical challenges that require highly specialised skills and interviewing protocols. Our objectives were to identify the most appropriate tools for collecting data from SSNs, key differences in interviewing students with and without special needs, and ethical and practical challenges during it. The ultimate goal was to develop ethical considerations and interviewing protocols for SSNs (ECIP4SSN). It was done through a two‐step Delphi method, collecting data from 11 special education experts from 11 different countries. The experts believed the tools should be selected based on the study's nature; however, many of them preferred the interviews. They believed that interviewing special students is challenging due to their special needs. Challenges pertaining to consent, hyperactivity, emotional distress, and wrong and irrelevant answers were among the top challenges mentioned by them. Final ECIP4SSN consisted of the key ethical considerations and interviewing protocols i.e. obtaining the background information and support, training and preparations, consent, data collection, building rapport, emotional support, confidentiality, and anonymity, using supporting tools, language, methods of asking questions and validating responses. The ECIP4SSN will guide the university teachers, students and researchers in interviewing special students.

Are primary care physicians familiar enough with potentially inappropriate medications in geriatric care? A cross-sectional study in the Eastern Province of Saudi Arabia.

Potentially inappropriate medications (PIMs) and polypharmacy constitute increasing healthcare costs and significant risk for adverse outcomes in older adults. The American Geriatrics Society Beers Criteria form a screening tool for the identification of PIMs and guidance for healthcare providers in prescribing appropriate medications. However, primary care physicians' knowledge of screening tools, of Beers Criteria, in particular, is not known. Therefore, this study was to investigate primary care physicians in the Eastern Province of Saudi Arabia and their awareness of Beers Criteria and knowledge of PIMs. This cross-sectional study was conducted among primary care physicians working in the Eastern Province of Saudi Arabia. Data were collected using an online self-administered questionnaire that consisted of sections on the general characteristics of respondents and their knowledge of Beers Criteria as a screening tool. Eight clinical-based vignettes concerning different therapeutic areas of medication use in the elderly were included, with a score of 1 and 0 for correct and wrong answers, respectively. Data presented as frequency and percentage. Chi-square test was used to determine the association between duration of practice and the level of awareness about Beers criteria. Of the 121 physicians who returned completed questionnaires, 41.3% of respondents knew about Beers Criteria. Most respondents (52.9%) were confident in prescribing appropriately for elderly patients. The association between the duration of practice and confidence level was statistically significant (P = 0.040). Respondents showed an above-average knowledge of the clinical vignettes with a correct answer rate >50% in all clinical scenarios. Online search (84.2%) and physician colleagues' knowledge and experiences (39.2%) were the primary source of information reported by the respondents. Awareness of Beers Criteria of primary care physicians in Saudi Arabia's Eastern Province is low. Therefore, our results will educate healthcare workers on the importance of Beers Criteria in Geriatric patients' prescriptions, in order to significantly improve the well-being of the elderly.

Albuterol off-label use in the management of hyperkalemia and students survey

This study aimed to explore the evidence in the off-label use of albuterol in the acute management of hyperkalemia. It also aims to explore the knowledge and opinion of first-year pharmacy students related to this topic. Based on the literature survey, it shows a lack of awareness among first-year pharmacy students regarding albuterol, hyperkalemia, and the use of albuterol for hyperkalemia. Demographic factors, such as income and years of work experience, were shown to influence students' understanding of the topic. Regarding the survey, a total of forty participants completed an online survey, providing information on demographic factors such as age, gender, residency state, employment status, income, education, and work experience. Results revealed a predominantly female and younger student population, with diverse geographic representation and varied employment and income levels. Most participants selected the wrong answer to questions related to albuterol and hyperkalemia, suggesting potential knowledge gaps in this area. These findings highlight the need for targeted educational interventions to enhance student understanding and clinical decision-making skills in pharmacotherapy-related concepts. Further research could evaluate the effectiveness of such interventions in improving student knowledge and patient care outcomes.

More haste, less speed? Relationship between response time and response accuracy in gamified online quizzes in an undergraduate engineering course

This study delves into the relationship between students' response time and response accuracy in Kahoot! quiz games, within the context of an app development course for first year university students. Previous research on response time in standardized tests has suggested that longer response time are often linked to correct answers, a phenomenon known as speed-accuracy trade-off. However, our study represents one of the initial investigations into this relationship within gamified online quizzes. Unlike standardized tests, each of our Kahoot! quiz modules consisted of only five single choice or true/false questions, with a time limit for each question and instantaneous feedback provided during game play. Drawing from data collected from 21 quiz modules spanning from 2022 to 2023, we compiled a dataset comprising 4640 response-time/accuracy entries. Mann-Whitney U tests revealed significant differences in normalized response time between correct answers and wrong answers in 17 out of the 21 quiz modules, with wrong answers exhibiting longer response time. The effect size of this phenomenon varied from modest to moderate. Additionally, we observed significant negative correlations between normalized overall response time and overall accuracy across all quiz modules, with only two exceptions. These findings challenge the notion of the speed-accuracy trade-off but align with some previous studies that have identified heterogeneity speed-accuracy relationships in standardized tests. We contextualized our findings within the framework of dual processing theory and underscored their implications to learning assessment in online quiz environments.

A Comparative Analysis of ChatGPT, ChatGPT-4, and Google Bard Performances at the Advanced Burn Life Support Exam.

Artificial intelligence and large language models (LLMs) have recently gained attention as promising tools in various healthcare domains, offering potential benefits in clinical decision-making, medical education, and research. The Advanced Burn Life Support (ABLS) program is a didactic initiative endorsed by the American Burn Association, aiming to provide knowledge on the immediate care of severely burn patients. The aim of the study was to compare the performance of three LLMs (ChatGPT-3.5, ChatGPT-4, and Google Bard) on the ABLS exam. The ABLS exam consists of 50 questions with 5 multiple-choice answers. The passing threshold is 80% of correct answers. The 3 LLMs were queried with the 50 questions included in the latest version of the ABLS exam, on July 18th, 2023. ChatGPT-3.5 scored 86% (43 out of 50), ChatGPT-4 scored 90% (45 out of 50), and Bard scored 70% (35 out of 50). No difference was measured between ChatGPT-3.5 and ChatGPT-4 (P = .538) and between ChatGPT-3.5 and Bard (P = .054), despite the borderline P-value. ChatGPT-4 performed significantly better than Bard (P = .012). Out of the 50 questions, 78% (n = 39) were direct questions, while 12% (n = 11) were presented as clinical scenarios. No difference in the rate of wrong answers was found based on the type of question for the 3 LLMs. ChatGPT-3.5 and ChatGPT-4 demonstrated high accuracy at the ABLS exam and outperformed Google Bard. However, the potential multiple applications of LLMs in emergency burn and trauma care necessitate appropriate surveillance and most likely should represent a tool to complement human cognition.

Performance of ChatGPT on American Board of Surgery In-Training Examination Preparation Questions

IntroductionChat Generative Pretrained Transformer (ChatGPT) is a large language model capable of generating human-like text. This study sought to evaluate ChatGPT’s performance on Surgical Council on Resident Education (SCORE) self-assessment questions. MethodsGeneral surgery multiple choice questions were randomly selected from the SCORE question bank. ChatGPT (GPT-3.5, April-May 2023) evaluated questions and responses were recorded. ResultsChatGPT correctly answered 123 of 200 questions (62%). ChatGPT scored lowest on biliary (2/8 questions correct, 25%), surgical critical care (3/10, 30%), general abdomen (1/3, 33%), and pancreas (1/3, 33%) topics. ChatGPT scored higher on biostatistics (4/4 correct, 100%), fluid/electrolytes/acid-base (4/4, 100%), and small intestine (8/9, 89%) questions. ChatGPT answered questions with thorough and structured support for its answers. It scored 56% on ethics questions and provided coherent explanations regarding end-of-life discussions, communication with coworkers and patients, and informed consent. For many questions answered incorrectly, ChatGPT provided cogent, yet factually incorrect descriptions, including anatomy and steps of operations. In two instances, it gave a correct explanation but chose the wrong answer. It did not answer two questions, stating it needed additional information to determine the next best step in treatment. ConclusionsChatGPT answered 62% of SCORE questions correctly. It performed better at questions requiring standard recall but struggled with higher-level questions that required complex clinical decision making, despite providing detailed responses behind its rationale. Due to its mediocre performance on this question set and sometimes confidently-worded, yet factually inaccurate responses, caution should be used when interpreting ChatGPT’s answers to general surgery questions.