دراسة قبول الطلاب واستخدامهم لـ ChatGPT في التعليم العالي السعودي
Examining Students’ Acceptance and Use of ChatGPT in Saudi Arabian Higher Education

المجلة: European Journal of Investigation in Health Psychology and Education، المجلد: 14، العدد: 3
DOI: https://doi.org/10.3390/ejihpe14030047
PMID: https://pubmed.ncbi.nlm.nih.gov/38534908
تاريخ النشر: 2024-03-17
المؤلف: Abu Elnasr E. Sobaih وآخرون
الموضوع الرئيسي: اعتماد التكنولوجيا وسلوك المستخدم

نظرة عامة

تستكشف هذه الدراسة قبول واستخدام ChatGPT بين الطلاب في التعليم العالي في المملكة العربية السعودية، مما يعكس الاهتمام المتزايد بهذه التكنولوجيا منذ إطلاقها في عام 2022. باستخدام استبيان ذاتي يعتمد على “النظرية الموحدة لقبول واستخدام التكنولوجيا” (UTAUT2)، تم جمع البيانات من 520 طالبًا في جامعة حكومية في بداية العام الدراسي 2023-2024. أظهرت النتائج من نمذجة المعادلات الهيكلية أن توقع الأداء (PE) والتأثير الاجتماعي (SI) وتوقع الجهد (EE) أثرت بشكل كبير على النية السلوكية (BI) لاستخدام ChatGPT، مما أثر بدوره على الاستخدام الفعلي. ومع ذلك، انحرفت الدراسة عن الأبحاث السابقة من خلال الكشف عن علاقة سلبية بين ظروف التيسير (FCs) وBI، بالإضافة إلى علاقة غير ذات دلالة بين FCs والاستخدام الفعلي، والتي تعزى إلى نقص الموارد والدعم الخارجي.

تخلص الدراسة إلى أنه بينما يتم دعم إطار عمل UTAUT جزئيًا، هناك حاجة ماسة لمؤسسات التعليم العالي لتوفير موارد ودعم كافيين للطلاب للاستفادة بشكل فعال من ChatGPT لأغراض أكاديمية. تسلط النتائج الضوء على أهمية تعزيز فهم قيود أدوات الذكاء الاصطناعي وتعزيز الاستخدام المسؤول لضمان نتائج تعلم مستدامة. يتم تشجيع المؤسسات على تسهيل بيئة تعزز التطبيق الأخلاقي لـ ChatGPT، مما يثري التجربة التعليمية للطلاب.

مقدمة

تناقش مقدمة ورقة البحث الاهتمام المتزايد في استخدام Chat Generative Pre-trained Transformer (ChatGPT) في التعليم العالي منذ إطلاقه في عام 2022. يُعترف بـ ChatGPT، الذي طورته OpenAI، لقدرته على مساعدة الطلاب في مختلف المهام الأكاديمية، بما في ذلك الواجبات وتطوير مهارات اللغة. بينما تسلط الدراسات الضوء على فوائد مثل الدعم الفوري وتعزيز تجارب التعلم، لا تزال المخاوف قائمة بشأن تأثيره على نزاهة الأكاديمية، وإمكانية المعلومات المنحازة، وخطر تقويض مهارات التفكير النقدي. يؤكد العلماء على الحاجة إلى الحذر لمنع الاعتماد المفرط على أدوات الذكاء الاصطناعي مثل ChatGPT في السياقات التعليمية.

تحدد الورقة فجوة بحثية في فهم العوامل التي تؤثر على قبول الطلاب واستخدامهم لـ ChatGPT في التعليم، لا سيما ضمن نظام التعليم العالي في المملكة العربية السعودية. استكشفت الدراسات السابقة مجموعة متنوعة من دوافع استخدام ChatGPT، مثل سهولة الاستخدام ومساعدة الواجبات، ولكن لا يزال هناك حاجة لتحليل شامل في سياقات ثقافية مختلفة. تهدف هذه الدراسة إلى سد هذه الفجوة من خلال تطبيق إطار عمل النظرية الموحدة لقبول واستخدام التكنولوجيا (UTAUT) لفحص المتغيرات التي تؤثر على تفاعل الطلاب السعوديين مع ChatGPT. من المتوقع أن توفر النتائج رؤى لقادة الجامعات والمعلمين لإدارة دمج أدوات الذكاء الاصطناعي بشكل أفضل في البيئات الأكاديمية. ستراجع الأقسام اللاحقة إطار عمل UTAUT، وتحدد منهجية البحث، وتقدم النتائج، وتقدم توصيات للدراسات المستقبلية.

الطرق

تحدد قسم “الطرق” في ورقة البحث التصميم التجريبي والتقنيات التحليلية المستخدمة للتحقيق في أسئلة البحث. استخدمت الدراسة نهجًا كميًا، مع دمج التحليلات الإحصائية لتقييم البيانات التي تم جمعها من تجارب مختلفة. تضمنت المنهجيات المحددة تجارب محكومة، واستبيانات، أو محاكاة، اعتمادًا على طبيعة مشكلة البحث.

شملت جمع البيانات إجراءات منهجية لضمان الموثوقية والصلاحية، مع تطبيق تقنيات أخذ عينات مناسبة لجمع عينات تمثيلية. تم إجراء التحليل باستخدام برامج إحصائية معتمدة، مما مكن الباحثين من إجراء حسابات معقدة واستنتاج رؤى ذات مغزى من البيانات. تم تعريف المقاييس والمتغيرات الرئيسية بوضوح، مما يسهل تفسيرًا قويًا للنتائج. بشكل عام، كانت الطرق المستخدمة مصممة لاختبار الفرضيات بدقة والمساهمة في فهم الموضوع قيد البحث.

النتائج

قيمت نتائج الدراسة الخصائص النفسية لمقاييس مختلفة من خلال النموذج الخارجي، مؤكدة صلاحية تقارب قوية مع تحميلات معيارية فوق 0.7، وتناسق داخلي كما تشير إليه α لـ Cronbach وموثوقيات مركبة (CRs) تتجاوز العتبة 0.7. تجاوز متوسط التباين المستخرج (AVE) لجميع البنى 0.5 الموصى بها، مما يؤكد صلاحية تقارب مرضية. تم تأسيس صلاحية التمييز باستخدام كل من طريقة فورنيل ولارك و نسبة Heterotrait-Monotrait (HTMT)، مع بقاء جميع قيم HTMT تحت 0.9، مما يؤكد صلاحية التمييز.

أشارت قيم R² المعاد حسابها إلى أن سهولة الإدراك (PE) وتوقع الجهد (EE) والتأثير الاجتماعي (SI) وظروف التيسير (FC) شكلت 30% من التباين في النية السلوكية (BI)، بينما أوضحت هذه البنى جنبًا إلى جنب مع BI 74% من التباين في استخدام ChatGPT. وُجدت تأثيرات إيجابية كبيرة لـ BI على PE (β = 0.398) وEE (β = 0.144) وSI (β = 0.445)، مما يدعم الفرضيات H1 وH2 وH3. على العكس، أثرت FC سلبًا على BI (β = -0.204)، مما أدى إلى رفض H4. بالإضافة إلى ذلك، كان لـ BI تأثير إيجابي قوي على استخدام ChatGPT (β = 0.789)، مما يدعم H9. كما كشفت التحليلات أن BI وسّطت العلاقات بين PE وEE وSI وFC مع استخدام ChatGPT، مما يدعم الفرضيات H10 وH11 وH12 وH13.

المناقشة

تسلط قسم المناقشة في ورقة البحث الضوء على قبول واستخدام ChatGPT بين طلاب التعليم العالي في المملكة العربية السعودية، باستخدام إطار عمل UTAUT لتحليل العوامل الرئيسية التي تؤثر على النوايا السلوكية (BIs) والاستخدام الفعلي. تشير النتائج إلى أن التأثير الاجتماعي (SI) هو أقوى مؤشر على نوايا الطلاب لاستخدام ChatGPT لأغراض أكاديمية، مما يعكس السياق الثقافي الجماعي حيث تشكل توصيات الأقران بشكل كبير المواقف تجاه اعتماد التكنولوجيا. كما ظهر توقع الأداء (PE) كعامل حاسم، حيث يرى الطلاب أن ChatGPT أداة قيمة تعزز إنتاجيتهم الأكاديمية وأدائهم. على العكس، وُجد أن توقع الجهد (EE) يؤثر إيجابيًا على BIs ولكنه لم يتحول إلى استخدام فعلي، مما يشير إلى أنه بينما يجد الطلاب ChatGPT سهل الاستخدام، فإن هذا الإدراك وحده لا يضمن دمجه في روتينهم الأكاديمي.

من المثير للاهتمام، أن ظروف التيسير (FC) أظهرت تأثيرًا سلبيًا على BIs، مما يشير إلى أن نقص الدعم المؤسسي والموارد يعيق استعداد الطلاب لتبني ChatGPT. يتعارض هذا مع إطار عمل UTAUT، الذي يفترض أن FC يجب أن تسهل قبول التكنولوجيا. تؤكد الدراسة على ضرورة أن تقدم مؤسسات التعليم العالي موارد ودعم كافيين لتعزيز الاستخدام الفعال لـ ChatGPT، إلى جانب تعزيز الوعي بكل من فوائد وقيود أدوات الذكاء الاصطناعي. تختتم البحث بالقول إنه بينما ينطبق إطار عمل UTAUT جزئيًا، فإنه يتطلب إعادة تقييم لكيفية تصور FC في سياق اعتماد التكنولوجيا في التعليم، مما يبرز أهمية الاستخدام المسؤول والأخلاقي لأدوات الذكاء الاصطناعي بين الطلاب.

القيود

تسلط القيود في هذه الدراسة، التي أجريت على طلاب إدارة في مؤسسة تعليم عالٍ عامة في جنوب أفريقيا، الضوء على عدة مجالات للبحث المستقبلي. بينما قدمت النتائج رؤى دعمت وتعارضت مع إطار عمل النظرية الموحدة لقبول واستخدام التكنولوجيا (UTAUT)، فإن نقص تنوع العينة – تحديدًا استبعاد المشاركين من المؤسسات الخاصة – يحد من تعميم النتائج. يجب أن تأخذ الأبحاث المستقبلية في الاعتبار تضمين مجموعة أوسع من السياقات التعليمية لفهم أفضل لتصورات أدوات الذكاء الاصطناعي مثل ChatGPT في التعليم.

بالإضافة إلى ذلك، لم تستكشف الدراسة التأثيرات المعدلة للمتغيرات الديموغرافية مثل الجنس والعمر والتخصص وسنة الدراسة والخبرة السابقة على اعتماد التكنولوجيا في البيئات التعليمية. يمكن أن يؤدي التحقيق في هذه العوامل إلى الحصول على رؤى قيمة حول تفاصيل قبول التكنولوجيا. يمكن أن تستفيد الدراسات المستقبلية أيضًا من مقارنة تجارب الطلاب الجامعيين وطلاب الدراسات العليا عبر دول مختلفة، مما يثري الفهم لدمج الذكاء الاصطناعي في بيئات تعليمية متنوعة.

Journal: European Journal of Investigation in Health Psychology and Education, Volume: 14, Issue: 3
DOI: https://doi.org/10.3390/ejihpe14030047
PMID: https://pubmed.ncbi.nlm.nih.gov/38534908
Publication Date: 2024-03-17
Author(s): Abu Elnasr E. Sobaih et al.
Primary Topic: Technology Adoption and User Behaviour

Overview

This study investigates the acceptance and utilization of ChatGPT among students in Saudi Arabian higher education, reflecting the increasing interest in this technology since its launch in 2022. Utilizing a self-report survey based on the “Unified Theory of Acceptance and Use of Technology” (UTAUT2), data were collected from 520 students at a public university at the beginning of the 2023-2024 academic year. The findings from structural equation modeling indicated that performance expectancy (PE), social influence (SI), and effort expectancy (EE) significantly influenced behavioral intention (BI) to use ChatGPT, which in turn affected actual usage. However, the study diverged from previous research by revealing a negative relationship between facilitating conditions (FCs) and BI, as well as an insignificant relationship between FCs and actual use, attributed to a lack of external resources and support.

The study concludes that while the UTAUT framework is partially supported, there is a critical need for higher education institutions to provide adequate resources and support for students to effectively leverage ChatGPT for academic purposes. The results highlight the importance of fostering an understanding of the limitations of AI tools and promoting responsible usage to ensure sustainable learning outcomes. Institutions are encouraged to facilitate an environment that enhances the ethical application of ChatGPT, thereby enriching the educational experience for students.

Introduction

The introduction of the research paper discusses the increasing interest in the use of Chat Generative Pre-trained Transformer (ChatGPT) in higher education since its launch in 2022. ChatGPT, developed by OpenAI, is recognized for its ability to assist students with various academic tasks, including assignments and language skills development. While studies highlight benefits such as prompt support and enhanced learning experiences, concerns persist regarding its impact on academic integrity, the potential for biased information, and the risk of undermining critical thinking skills. Scholars emphasize the need for caution to prevent over-dependence on AI tools like ChatGPT in educational contexts.

The paper identifies a research gap in understanding the factors influencing students’ acceptance and use of ChatGPT in education, particularly within Saudi Arabia’s higher education system. Previous studies have explored various drivers of ChatGPT usage, such as ease of use and homework assistance, but there remains a need for comprehensive analysis in different cultural contexts. This research aims to fill this gap by applying the Unified Theory of Acceptance and Use of Technology (UTAUT) framework to examine the variables affecting Saudi students’ engagement with ChatGPT. The findings are expected to provide insights for university leaders and educators to better manage the integration of AI tools in academic settings. The subsequent sections will review the UTAUT framework, outline the research methodology, present results, and offer recommendations for future studies.

Methods

The “Methods” section of the research paper outlines the experimental design and analytical techniques employed to investigate the research questions. The study utilized a quantitative approach, incorporating statistical analyses to evaluate the data collected from various experiments. Specific methodologies included controlled experiments, surveys, or simulations, depending on the nature of the research problem.

Data collection involved systematic procedures to ensure reliability and validity, with appropriate sampling techniques applied to gather representative samples. The analysis was conducted using established statistical software, enabling the researchers to perform complex calculations and derive meaningful insights from the data. Key metrics and variables were defined clearly, facilitating a robust interpretation of the results. Overall, the methods employed were designed to rigorously test the hypotheses and contribute to the field’s understanding of the topic under investigation.

Results

The results of the study assessed the psychometric properties of various scales through the outer model, confirming robust convergent validity with standardized loadings above 0.7, and internal consistency as indicated by Cronbach’s α and composite reliabilities (CRs) exceeding the threshold of 0.7. The average variance extracted (AVE) for all constructs surpassed the recommended 0.5, affirming satisfactory convergent validity. Discriminant validity was established using both Fornell and Larcker’s method and the Heterotrait-Monotrait (HTMT) ratio, with all HTMT values remaining below 0.9, thus confirming discriminant validity.

The bootstrapped R² values indicated that perceived ease (PE), effort expectancy (EE), social influence (SI), and facilitating conditions (FC) accounted for 30% of the variance in behavioral intention (BI), while these constructs along with BI explained 74% of the variance in ChatGPT usage. Significant positive effects were found for BI on PE (β = 0.398), EE (β = 0.144), and SI (β = 0.445), supporting hypotheses H1, H2, and H3. Conversely, FC negatively impacted BI (β = -0.204), leading to the rejection of H4. Additionally, BI had a strong positive effect on ChatGPT usage (β = 0.789), supporting H9. The analysis also revealed that BI mediated the relationships between PE, EE, SI, and FC with ChatGPT usage, supporting hypotheses H10, H11, H12, and H13.

Discussion

The discussion section of the research paper highlights the acceptance and usage of ChatGPT among higher education students in Saudi Arabia, utilizing the UTAUT framework to analyze key determinants influencing behavioral intentions (BIs) and actual usage. The findings indicate that social influence (SI) is the strongest predictor of students’ intentions to use ChatGPT for academic purposes, reflecting the collective cultural context where peer recommendations significantly shape attitudes towards technology adoption. Performance expectancy (PE) also emerged as a critical factor, with students perceiving ChatGPT as a valuable tool that enhances their academic productivity and performance. Conversely, effort expectancy (EE) was found to positively influence BIs but did not translate into actual usage, suggesting that while students find ChatGPT user-friendly, this perception alone does not guarantee its integration into their academic routines.

Interestingly, facilitating conditions (FC) were shown to negatively impact BIs, indicating that a lack of institutional support and resources hinders students’ willingness to adopt ChatGPT. This contradicts the UTAUT framework, which posits that FC should facilitate technology acceptance. The study underscores the necessity for higher education institutions to provide adequate resources and support to foster effective use of ChatGPT, alongside promoting awareness of both the benefits and limitations of AI tools. The research concludes that while the UTAUT framework partially applies, it calls for a reevaluation of how FC is perceived in the context of technology adoption in education, emphasizing the importance of responsible and ethical use of AI tools among students.

Limitations

The limitations of this study, conducted on management students at a public higher education institution in South Africa, highlight several areas for future research. While the findings provided insights that both supported and contradicted the Unified Theory of Acceptance and Use of Technology (UTAUT) framework, the sample’s lack of diversity—specifically, the exclusion of participants from private institutions—limits the generalizability of the results. Future research should consider including a broader range of educational contexts to better understand perceptions of AI tools like ChatGPT in education.

Additionally, the study did not explore the moderating effects of demographic variables such as gender, age, specialization, study year, and prior experience on technology adoption in educational settings. Investigating these factors could yield valuable insights into the nuances of technology acceptance. Future studies could also benefit from comparing the experiences of undergraduate and postgraduate students across different countries, thereby enriching the understanding of AI integration in diverse educational environments.