الذكاء الاصطناعي وتكنولوجيا الاتصال في الأوساط الأكاديمية: تصورات أعضاء هيئة التدريس واعتماد الذكاء الاصطناعي التوليدي Artificial intelligence and communication technologies in academia: faculty perceptions and the adoption of generative AI

المجلة: International Journal of Educational Technology in Higher Education، المجلد: 22، العدد: 1
DOI: https://doi.org/10.1186/s41239-025-00511-7
تاريخ النشر: 2025-03-13

الذكاء الاصطناعي وتكنولوجيا الاتصال في الأوساط الأكاديمية: تصورات أعضاء هيئة التدريس واعتماد الذكاء الاصطناعي التوليدي

آية شطة و كيندال هارتلي²(D)

*المراسلة:
aya.shata@unlv.edu
مدرسة هانك غرينسبون للصحافة ودراسات الإعلام، جامعة نيفادا، لاس فيغاس، 4505 S. Maryland Parkway، صندوق #45007، لاس فيغاس، NV 89154-5007، الولايات المتحدة الأمريكية
قسم التعليم والتعلم، جامعة نيفادا، لاس فيغاس، 4505 S. Maryland Parkway، صندوق #453005، لاس فيغاس، NV 89154-3005، الولايات المتحدة الأمريكية

الملخص

الذكاء الاصطناعي (AI) يفتح عصرًا من التحول المحتمل في مجالات متعددة، خاصة في تكنولوجيا الاتصال التعليمية، مع أدوات مثل ChatGPT وغيرها من تطبيقات الذكاء الاصطناعي التوليدي (GenAI). هذا الانتشار السريع واعتماد أدوات GenAI أثار اهتمامًا وقلقًا كبيرين بين أساتذة الجامعات، الذين يتعاملون مع الديناميات المتطورة في الاتصال الرقمي داخل الفصول الدراسية. ومع ذلك، فإن تأثير ونتائج GenAI في التعليم لا تزال غير مدروسة بشكل كاف. لذلك، تستخدم هذه الدراسة نموذج قبول التكنولوجيا (TAM) ونظرية التعلم الاجتماعي (SCT) كإطارات نظرية لاستكشاف تصورات أعضاء هيئة التدريس في التعليم العالي، والمواقف، والاستخدام، والدوافع، كعوامل أساسية تؤثر على اعتمادهم أو رفضهم لأدوات GenAI. تم إجراء استبيان بين أعضاء هيئة التدريس بدوام كامل في التعليم العالي ( ) تم تجنيدهم من جامعتين حكوميتين متوسطتي الحجم في الولايات المتحدة. وجدت النتائج أن الفائدة المدركة للذكاء الاصطناعي من قبل أساتذة الجامعات تنبأت بمواقفهم ونواياهم لاستخدام واعتماد التكنولوجيا، أكثر من سهولة الاستخدام المدركة. الثقة والتعزيز الاجتماعي أثرت بشكل كبير على قرارات اعتماد GenAI من قبل أساتذة الجامعات وعملت كوسيطين مهمين لفهم العلاقة بين TAM وSCT بشكل أفضل. أكدت النتائج على قوة الديناميات الاجتماعية في تشكيل كفاءة الأساتذة الذاتية، والمواقف، واستخدام GenAI. تعزز الثقة تأثير الأقران وتؤثر على كيفية تشكيل الفائدة المدركة لاستعداد المستخدمين لاعتماد التكنولوجيا، بينما كان لكفاءة الذات تأثير ضئيل. توفر هذه الدراسة رؤى قيمة تُعلم سياسات التعليم العالي التي تهدف إلى تحسين تجربة التعليم للطلاب الجامعيين في قوة العمل المدفوعة بالذكاء الاصطناعي.

الكلمات المفتاحية: الذكاء الاصطناعي، تكنولوجيا الاتصال، التعليم العالي، اعتماد التكنولوجيا، أعضاء هيئة التدريس، الثقة، كفاءة الذات، التأثير الاجتماعي

المقدمة

أصبح الذكاء الاصطناعي التوليدي (GenAI) سريعًا ذا أهمية خاصة في التعليم العالي. تشير الأبحاث الحديثة إلى أن من البالغين الذين تتراوح أعمارهم بين قد استخدموا ChatGPT (McClain، 2024). وهذا ارتفاع من قبل تسعة أشهر فقط. يواجه أعضاء هيئة التدريس في التعليم العالي ديناميكية فصل دراسي تتغير بسرعة حيث تثير هذه الأدوات متعددة الوسائط المستخدمة على نطاق واسع العديد من الفرص والتحديات. تستكشف هذه الدراسة هذه الفرص والتحديات من خلال استكشاف الاستخدام الحالي ووجهات نظر أعضاء هيئة التدريس في التعليم العالي
بشأن GenAI، أحدث الإضافات إلى مجموعة أدوات تكنولوجيا الاتصال والإعلام التعليمية.
يُعرف الذكاء الاصطناعي التوليدي (GenAI) بأنه تقنية تستخدم نماذج التعلم العميق لإنشاء محتوى يشبه المحتوى البشري (مثل الصور، الكلمات) استجابةً للمطالبات (Lim et al.، 2023). تشمل أمثلة GenAI التفاعلية المعتمدة على النص ChatGPT من OpenAI، وGemini من Google، وClaude من Anthropic، وMeta.AI. كما أن أدوات GenAI متعددة الوسائط أصبحت أكثر انتشارًا وتغير المشهد في إنشاء الوسائط التعليمية. يمكن استخدام أدوات تطوير الصور مثل DALL-E أو Midjourney لإنشاء محتوى بصري. يتقدم تطوير الفيديو بسرعة مع أدوات مثل Sora من Open AI وRunwayML. معًا، فإن الإمكانية لإنشاء مواد تعليمية رقمية أكثر جاذبية وتنوعًا كبيرة.
مثل معظم المجالات، تظهر التطبيقات في سياق التعليم العالي بسرعة، وتزيد القدرات المتزايدة من صعوبة تحديد أكثر حالات الاستخدام بروزًا. تشمل أمثلة التدريس تطوير دراسات حالة وعناصر تقييم لموضوع معين. تشمل أمثلة البحث الدعم لتحليل البيانات النوعية. يشمل الدعم اللوجستي القدرة على إنشاء تسميات توضيحية للصور ومقاطع الفيديو. كما قدمت أحدث نماذج LLM قفزة مذهلة في دقة تحويل الكلام إلى نص.
بينما ليس الذكاء الاصطناعي ظاهرة جديدة وقد تم دراسته لسنوات (عقود في بعض المجالات)، فإن الانتشار السريع واعتماد أدوات GenAI أثار قلقًا واهتمامًا واسع النطاق بين المعلمين والباحثين (Yusuf et al.، 2024). لقد انتقل هذا الاهتمام بتداعيات الذكاء الاصطناعي إلى ما وراء التخصصات الأكاديمية الأصلية مثل علوم الكمبيوتر، وإدارة المعلومات، وغيرها من مجالات STEM. الآن، تحظى التداعيات لتكنولوجيا الاتصال والتعليم خارج تخصصات STEM الأكاديمية بمزيد من الاهتمام وهي هدف هذه الدراسة. تركيزنا هو على التداعيات في المجالات غير STEM حيث من المحتمل أن لا يكون لدى الأكاديميين فهم عميق للتكنولوجيا الأساسية. بينما يحد هذا من إمكانية تعميم النتائج، فإنه يوفر رؤى قيمة حول التأثير الناشئ للذكاء الاصطناعي على التخصصات التي كانت تقليديًا أقل تأثرًا بمثل هذه التقنيات.

مراجعة الأدبيات

الذكاء الاصطناعي في التعليم العالي

الذكاء الاصطناعي ليس ظاهرة جديدة؛ فقد تم دراسة تطبيقاته في التعليم العالي لعقود (Hartley et al.، 2024). لقد حسنت التقدمات الأخيرة في التكنولوجيا الأساسية لتطبيقات الذكاء الاصطناعي بشكل كبير من القدرات العامة (Mollick، 2024). قدمت هذه التقدمات العديد من التحديات والفرص الجديدة لأعضاء هيئة التدريس في التعليم العالي (Yusuf et al.، 2024).
دور GenAI في الأوساط الأكاديمية وتداعياته على أعضاء هيئة التدريس تتطور وتصبح معقدة. يصف أحد الباحثين استخدام GenAI بأنه ‘ذكاء مشترك’ (Mollick، 2024)، وهو ما يتماشى مع التحذير من أن هذه الأدوات الجديدة يُفضل تصورها كمساعدين أو متعاونين. كما حدد الباحثون تناقضاتها، حيث يمكن اعتبارها داعمة ومعادية، فعالة وغير دقيقة، شائعة وممنوعة (Lim et al.، 2023). تتقاطع المهام الأكاديمية مثل التدريس، والخدمة، والبحث مع GenAI. وقد أظهرت الأبحاث المبكرة التي استخدمت GenAI لإنجاز مهام مثل تقديم ملاحظات الكتابة نتائج أقل مقارنةً بالمقيمين البشريين (Steiss et al.، 2024). ومع ذلك، فإن الأدوات
تتحسن بسرعة كما يتحسن فهمنا لكيفية ومتى نستخدم ChatGPT. الدراسة المذكورة أعلاه استخدمت ChatGPT الإصدار 3.5. وجدت دراسة تقارن بين ChatGPT الإصدار 3.5 و4.0 في امتحان الترخيص الطبي الأمريكي تحسنًا ملحوظًا. أظهر الإصدار 3.5 دقة قدرها ، بينما قفز الإصدار 4.0 إلى (Shieh et al.، 2024). ما إذا كان هذا التحسن يترجم إلى درجات أعلى للطلاب لا يزال غير واضح.

تصورات أعضاء هيئة التدريس والطلاب حول الذكاء الاصطناعي

جادل البعض بأن مسؤولية إعداد الطلاب للمستقبل تقتضي التحضير لـ ‘مجتمع مدفوع بالذكاء الاصطناعي’ (Chiu، 2024). لقد قدمت مجموعة متزايدة بسرعة من الأبحاث رؤى أولية حول أهمية GenAI في التعليم العالي. في مقارنة بين تصورات أعضاء هيئة التدريس والطلاب حول الكتابة باستخدام أدوات الذكاء الاصطناعي، وجد الباحثون اختلافات بشأن الاستخدام المقبول لـ GenAI (Barrett & Pack، 2023). ومن المثير للاهتمام، أن هذه الاختلافات أشارت إلى وجهة نظر أكثر إيجابية لـ GenAI من قبل المعلمين مقارنةً بالطلاب. تشير الأعمال التي سبقت إدخال ChatGPT إلى أن الآراء الإيجابية حول GenAI للتعلم قد تكون مبررة. في تحليل شامل لاستخدام الدردشة النصية في التعليم، وجد الباحثون العديد من المزايا للطلاب (Labadze et al.، 2023). وقد دعمت الأعمال الأكثر حداثة هذا الاكتشاف بتحسين الأداء على المدى القصير، ومع ذلك، هناك أيضًا أسباب للاعتقاد بأن استخدام GenAI يُفضل أن يُنظر إليه على أنه ‘عكاز’ يمكن أن يؤدي إزالته إلى انخفاض الفهم على المدى الطويل (Lim et al.، 2023). إحدى المخاوف الشائعة بين المعلمين هي إمكانية أن يؤدي GenAI إلى تراجع في مهارات الطلاب المعرفية والمنطقية (Yusuf et al.، 2024).

اعتماد التكنولوجيا

لفهم التفاعل بين أعضاء هيئة التدريس في التعليم العالي وGenAI، من المفيد النظر في أطر اعتماد التكنولوجيا. يقترح نموذج قبول التكنولوجيا (TAM) أن قبول التكنولوجيا يتأثر بشدة بمدى سهولة استخدامها المدركة (PEU) وفائدتها المدركة (PU) (ديفيس، 1989؛ غرانيتش ومارانغونيك، 2019؛ فينكاتيش وآخرون، 2003). لذلك، فإن تصورات أعضاء هيئة التدريس حول سهولة الاستخدام والفائدة هي طرق أساسية لاستكشاف آثار GenAI على التعليم العالي. إطار عمل آخر ذي صلة هو النظرية الموحدة لقبول واستخدام التكنولوجيا (UTAUT)، التي تعترف بقيمة مفاهيم TAM، وتضيف التأثير الاجتماعي والظروف الميسرة لدفع النية السلوكية لاستخدام التكنولوجيا (فينكاتيش وآخرون، 2016). لاستكشاف العوامل الاجتماعية والشخصية، تعتبر نظرية التعلم الاجتماعي المعرفي (SCT) إطارًا نظريًا آخر يشرح عملية قبول التكنولوجيا (باندورا، 1999). في هذه الدراسة، نجمع بين TAM وSCT لفهم أفضل لاستخدامات أساتذة الجامعات لـ GenAI.
باختصار، ستكون النية السلوكية لاستخدام التكنولوجيا مرتبطة مباشرة بالموقف تجاه التكنولوجيا، والذي يتأثر بدوره بمدى سهولة الاستخدام المدركة (PEU) ومدى الفائدة المدركة (PU) للتكنولوجيا. بالإضافة إلى ذلك، ستساهم عوامل أخرى تم مناقشتها أدناه في معادلة الاعتماد.

الموقف

يمكن أن يؤثر موقف المستخدم تجاه التكنولوجيا والابتكار بشكل كبير على اعتماد تقنيات جديدة. تعتبر مواقف المستخدمين تجاه التقدم التكنولوجي
حيوية في تحديد ما إذا كانوا سيقبلون الابتكارات أم لا (ديفيس، 1989؛ غرانيتش ومارانغونيك، 2019). فيما يتعلق بالذكاء الاصطناعي، يمكن أن تؤثر مواقف الأفراد ومستويات راحتهم بشكل كبير على القبول والاندماج في مختلف جوانب نظام التعليم (باور وآخرون، 2024). تشير الدراسات الأولية إلى أن مواقف أعضاء هيئة التدريس تجاه Gen AI مختلطة (باور وآخرون، 2024). بينما يدرك غالبية أعضاء هيئة التدريس الإمكانية للتأثير الكبير، فإن هؤلاء الأساتذة أنفسهم أكثر دراية بالتكنولوجيا.

PEU/PU

يمكن أن يتأثر موقف الفرد تجاه التكنولوجيا بمدى سهولة استخدامها المدركة ومدى فائدتها المدركة (ديفيس وآخرون، 1989؛ غرانيتش ومارانغونيك، 2019). تصبح هذه العوامل حاسمة حيث يفكر أعضاء هيئة التدريس في التعليم العالي في اعتماد ودمج تقنيات الاتصال.
كان وصول ChatGPT في نوفمبر 2022 مهمًا جزئيًا لأنه قدم وسيلة سهلة للناس للتفاعل مع نموذج لغوي كبير متقدم. قبل ذلك، كان مستخدمو تطبيقات الذكاء الاصطناعي التوليدي محدودين إلى حد كبير لأولئك الذين لديهم معرفة بالبرمجة أو خبرة محددة في المجال. ومع ذلك، فإن سهولة الاستخدام المدركة محدودة إلى حد ما حيث كانت الواجهات المبكرة تعتمد بشكل أساسي على النص. مع تطور الواجهات لتتناسب بشكل أفضل مع احتياجات المستخدمين، تحسنت التصورات وستتحسن. على سبيل المثال، يمكن الآن لـ ChatGPT وClaude وGemini عرض نوافذ معاينة تتضمن مسودات عمل للصور والرسوم البيانية والجداول أو الشيفرات، مما يحسن بشكل كبير من سهولة الاستخدام (غوبتا وآخرون، 2024).

الكفاءة الذاتية

عند النظر في المواقف والتصورات تجاه التكنولوجيا الجديدة، فإن إطار عمل مفيد هو نظرية باندورا الاجتماعية المعرفية، التي تفترض أهمية الكفاءة الذاتية المحددة بالمحتوى في السعي وراء المعرفة (باندورا، 2002؛ فينكاتيش وآخرون، 2003). تشير الكفاءة الذاتية إلى اعتقاد الفرد في قدرته على النجاح في سياق أو مجال محتوى معين. يمكن أن تلعب الكفاءة الذاتية التكنولوجية دورًا وسطيًا أساسيًا في قبول التكنولوجيا (غرانيتش ومارانغونيك، 2019؛ هولدن ورادا، 2011). على سبيل المثال، ارتبطت سهولة الاستخدام المدركة بالكفاءة الذاتية لدى الطلاب الجامعيين المشاركين في التعلم الذاتي المدعوم بالتكنولوجيا (بان، 2020). من المحتمل أن تتأثر النية لاستخدام واعتماد أدوات GenAI بشكل مشابه بالكفاءة الذاتية لهذه التقنيات الناشئة. سيكون المستخدمون الذين خصصوا وقتًا كبيرًا لاستخدامها أكثر احتمالًا لاستخدام GenAI بثقة مقارنةً بأولئك الذين كانت مشاركتهم محدودة.

التعزيز الاجتماعي

تناولت الأبحاث أيضًا الدور المهم للتعزيز الاجتماعي. يمكن تعريف ذلك على أنه الدرجة التي يعتقد فيها المستخدمون أن شبكاتهم الاجتماعية تشعر أنه من المهم أن يعتمدوا تقنية معينة (ماركيان وآخرون، 2023). بينما خلصت تحليل شامل حديث إلى أن التأثير الاجتماعي المرتبط بالسلوك كان غير ذي دلالة، إلا أنها وجدت تأثيرًا كبيرًا على المواقف بشكل عام (ماركيان وآخرون، 2023). أدت السرعة في اعتماد GenAI، والاستثمارات الكبيرة من الشركات الكبرى، والتغطية الإعلامية الواسعة إلى ضغط مباشر وغير مباشر كبير على الأفراد نحو الاعتماد. بالطبع، تختلف هذه التأثيرات بين الثقافات والتخصصات. على سبيل المثال، لقد حظي تأثير GenAI على دورات الكتابة الجامعية باهتمام سلبي كبير.
تعترف نظرية التعلم الاجتماعي المعرفي بقدرتنا المعززة على التعلم بالملاحظة والدور الحيوي للتفاعلات الاجتماعية (باندورا، 1999). من المحتمل أن تؤثر هذه القدرة، جنبًا إلى جنب مع النقاش الواسع، على تصورات المستخدمين ومواقفهم وكفاءتهم الذاتية تجاه استخدام GenAI. بما يتماشى مع الأعمال السابقة في اعتماد التكنولوجيا (فينكاتيش وآخرون، 2016)، نقوم بتشغيل التأثير الاجتماعي كتعزيز اجتماعي، مما يعكس تأثير آراء وسلوكيات الآخرين على قرار الفرد باستخدام التكنولوجيا.

الثقة

تعتبر الثقة اعتبارًا رئيسيًا عند التحقيق في اعتماد تقنيات جديدة. يمكن وصف الثقة بأنها استعداد الفرد لقبول المعلومات من شخص آخر، مع توقع أن يكون مزود المعلومات دقيقًا (وو وآخرون، 2011). حدد تحليل شامل لدراسات اعتماد التكنولوجيا الثقة كعامل مرتبط بشدة بـ PEU وPU (ماركيان وآخرون، 2023). على سبيل المثال، استكشف تشونغ وآخرون (2022) دور الثقة في اعتماد تكنولوجيا مساعد الصوت الذكي، ووجدوا أن الثقة أثرت بشكل كبير على كل من الفائدة المدركة لمساعد الصوت الذكي ومواقف المستخدمين تجاهه، مما زاد من نية الاستخدام. وبالمثل، فحص تانانتونغ وونغراس (2024) استخدام الذكاء الاصطناعي في التوظيف من قبل الموارد البشرية ووجدوا أنه بينما لا تؤثر الثقة على النية بشكل مباشر، إلا أنها أثرت بشكل كبير على كل من الفائدة المدركة وسهولة الاستخدام. تؤكد هذه النتائج أن الثقة لها تأثير كبير على سهولة الاستخدام المدركة، والفائدة المدركة، والمواقف تجاه GenAI.
تعتبر الشبكات الاجتماعية والثقة جزءًا لا يتجزأ من انتشار التقنيات الجديدة (بوسكين، 2020). وفقًا لبوسكين، بمجرد أن يدرك الأفراد الفوائد المحتملة لتكنولوجيا ما، من الضروري بناء الثقة بأن الابتكار سيوفر المزايا الموعودة. لتحقيق ذلك، يعد التعزيز المستمر من الدوائر الاجتماعية أمرًا أساسيًا لتطوير الثقة في التكنولوجيا. وهذا يشير إلى أن التعزيز الاجتماعي يمكن أن يؤثر على تصورات الثقة التي بدورها تؤثر على احتمالية اعتماد التكنولوجيا، لا سيما في سياق GenAI. بشكل عام، تصبح الثقة مهمة بشكل فريد عندما يتعلق الأمر باعتماد GenAI، نظرًا للقلق الواسع حول الميل نحو الهلوسة في النماذج الرئيسية.

الدراسة الحالية

الغرض من هذه الدراسة هو تحسين فهمنا لتصورات واستخدامات ومواقف أعضاء هيئة التدريس في التعليم العالي تجاه GenAI، مع التركيز على تحديد الدوافع والحواجز من خلال إطار البحث الحالي ونظريات اعتماد التكنولوجيا. على وجه الخصوص، نستكشف العوامل التكنولوجية والاجتماعية والنفسية التي تم إثبات أنها ذات صلة. تشمل هذه العوامل تلك التي تم تحديدها في TAM وSCT بالإضافة إلى تصورات الثقة (الشكل 1).

أسئلة البحث المقترحة والافتراضات

RQ1: هل يعمل نموذج TAM بشكل مشابه مع GenAI كما هو الحال مع تقنيات أخرى؟
H1 (أ) ستؤثر الفائدة المدركة و(ب) سهولة الاستخدام المدركة بشكل إيجابي على مواقف أساتذة الجامعات تجاه GenAI.
الشكل 1 الإطار المفاهيمي وملخص الافتراضات
H2 (أ) الفائدة المدركة، (ب) سهولة الاستخدام المدركة، و(ج) مواقف أساتذة الجامعات ستؤثر إيجابياً على النية السلوكية لاستخدام الذكاء الاصطناعي التوليدي.
ستؤثر سهولة الاستخدام المدركة بشكل إيجابي على الفائدة المدركة لأساتذة الجامعات من استخدام الذكاء الاصطناعي التوليدي.
ستتوسط مواقف أساتذة الجامعات من الفئة H4 العلاقة بين (أ) الفائدة المدركة و(ب) سهولة الاستخدام المدركة على نيتهم السلوكية لاستخدام الذكاء الاصطناعي التوليدي.
RQ2: ما هي العلاقة بين نموذج TAM (المنفعة المتصورة والسهولة المتصورة) ونموذج SCT (الكفاءة الذاتية والتعزيز الاجتماعي) في فهم اعتماد الذكاء الاصطناعي التوليدي؟
تؤثر كفاءة الذات في الذكاء الاصطناعي التوليدي بشكل إيجابي على (أ) سهولة الاستخدام المدركة، (ب) مواقف أساتذة الجامعات، و(ج) النية السلوكية لاستخدام الذكاء الاصطناعي التوليدي.
H6 التعزيز الاجتماعي يؤثر بشكل إيجابي على (أ) الفائدة المدركة، (ب) سهولة الاستخدام المدركة، (ج) مواقف أساتذة الجامعات، و(د) الكفاءة الذاتية تجاه الذكاء الاصطناعي التوليدي.
H7 ستتوسط التعزيز الاجتماعي العلاقة بين (أ) سهولة الاستخدام المدركة، و (ب) الفائدة المدركة في نوايا أساتذة الجامعات السلوكية لاستخدام الذكاء الاصطناعي التوليدي.
RQ3: كيف يؤثر الثقة على العلاقة بين نموذج TAM (المنفعة المتوقعة وسهولة الاستخدام المتوقعة) ونموذج SCT (الكفاءة الذاتية والتعزيز الاجتماعي) في فهم اعتماد الذكاء العام؟
ثقة H8 في الذكاء الاصطناعي العام تؤثر بشكل إيجابي على (أ) الفائدة المدركة، (ب) سهولة الاستخدام المدركة، (ج) المواقف، و(د) نية السلوك تجاه الذكاء الاصطناعي العام.
ستقوم ثقة H9 بتعديل العلاقة بين الفائدة المدركة ونوايا أساتذة الجامعات السلوكية لاستخدام الذكاء الاصطناعي التوليدي.
ستتوسط ثقة H10 العلاقة بين التعزيز الاجتماعي ونوايا سلوك أساتذة الجامعات لاستخدام الذكاء الاصطناعي التوليدي.

طرق

استخدمت هذه الدراسة تصميم استبيان لاستكشاف تصورات أعضاء هيئة التدريس في التعليم العالي ومواقفهم واستخداماتهم وتجاربهم المتعلقة بالذكاء الاصطناعي التوليدي. كان الهدف الرئيسي هو فحص الدوافع المحتملة والعوائق والعوامل الكامنة وراء اعتماد أو رفض أدوات الذكاء الاصطناعي التوليدي بين أساتذة الجامعات.

المشاركون

تم تصميم استبيان عبر الإنترنت باستخدام Qualtrics مع أعضاء هيئة التدريس بدوام كامل في التعليم العالي تم تجنيدهم من جامعتين حكوميتين متوسطتي الحجم في الولايات المتحدة، واحدة على الساحل الشرقي والأخرى في الجنوب الغربي، للمشاركة في هذه الدراسة. مثلت العينة جميع أعضاء هيئة التدريس في مجالات العلوم الاجتماعية والإنسانية حيث يستخدمون GenAI بطريقة مشابهة تختلف جوهريًا عن تطبيقه في مجالات العلوم والتكنولوجيا والهندسة والرياضيات.

الإجراءات

بدأت إجراءات الدراسة بإنشاء قائمة بريد إلكتروني لجميع أعضاء هيئة التدريس المستمدة من دليل البريد الإلكتروني المتاح على الإنترنت، ثم تم تصفيتها حسب التخصص لتشمل فقط العلوم الاجتماعية والإنسانية. باستخدام توزيع البريد الإلكتروني عبر Qualtrics، أرسلنا بريدًا إلكترونيًا يدعوهم للمشاركة في الدراسة مع رابط للاستبيان. عند النقر على الرابط، يبدأون بإعطاء موافقتهم والإجابة على سؤال تصفية حول ما إذا كانوا قد استخدموا GenAI من قبل؛ إذا قالوا لا، نطرح أسئلة حول السبب، والانطباعات الحالية، والقلق بشأن الذكاء الاصطناعي، وأي نية مستقبلية، تليها بعض الأسئلة الديموغرافية ثم يتم إنهاء الاستبيان. إذا قالوا نعم، أنهم قد استخدموا GenAI، يستمرون إلى الاستبيان الرئيسي، حيث يبدأون بالإبلاغ عن راحتهم مع التكنولوجيا، واستخدامهم لتكنولوجيا GenAI في التدريس، والبحث، والخدمة، وآرائهم وقلقهم بشأن GenAI. ثم تناولت المفاهيم النظرية الرئيسية لنموذج قبول التكنولوجيا، تليها بعض الأسئلة الديموغرافية. كما تم إضافة بعض أسئلة التحقق من الانتباه إلى الاستبيان لضمان أن المشاركين يولون اهتمامًا كافيًا وتجنب الإسراع في الاستبيان أو إعطاء إجابات عشوائية قد تضر بجودة البيانات. تم جمع البيانات من يناير إلى مارس 2024.

تدابير

جميع البنى النظرية في هذه الدراسة استخدمت مقاييس موجودة، تم تعديلها لتناسب سياق الذكاء الاصطناعي وتم قياسها باستخدام مقاييس متعددة العناصر تم التحقق من صحتها في أبحاث سابقة.
تم قياس نية السلوك، وهي مستوى النية لتبني الذكاء الاصطناعي العام، باستخدام مقياس ليكرت من خمس نقاط تم تكييفه من (يوك وبارك، 2023) مع أربعة بيانات مثل “سأكون مستعدًا لاستخدام تقنية الذكاء الاصطناعي في المستقبل”. .
فائدة الاستخدام المدركة. الدرجة التي يعتقد فيها الشخص أن استخدام تكنولوجيا GenAI ستحسن أدائه في العمل (ديفيس وآخرون، 2024). تم قياس هذا البناء باستخدام مقياس ليكرت من خمس نقاط تم تكييفه من كل من (هولدن ورادا، 2011؛ يوك وبارك، 2023) مع سبع عبارات مثل “استخدام تكنولوجيا Gen. يحسن أدائي في عملي”. ).
سهولة الاستخدام المدركة، هي الدرجة التي يعتقد فيها الشخص أن استخدام GenAI سيكون خالياً من الجهد، تم قياسها باستخدام مقياس ليكرت المكون من تسع عناصر وخمس نقاط، تم تكييفه من (Holden & Rada، 2011) بما في ذلك عبارات مثل “أجد أنه من السهل جعل التكنولوجيا تفعل ما أريدها أن تفعله”. .
الكفاءة الذاتية للذكاء الاصطناعي هي ثقة المستخدمين الشخصية في استخدام تقنية الذكاء الاصطناعي بنجاح وهدف. تم اعتماد المقياس من (هولدن ورادا، 2011). استخدم البناء مقياس جوتمان المكون من عشرة نقاط. غير واثق على الإطلاق في واثق تمامًا) مثل “يمكنني إكمال أي مهمة مرغوبة باستخدام تقنية GenAI إذا لم يكن هناك أحد حولي ليخبرني بما يجب القيام به أثناء ذلك.” .
المواقف تجاه الذكاء الاصطناعي العام تشير إلى التقييم العاطفي والمعرفي الشامل للفرد تجاه تقنية جديدة. تم قياس ذلك باستخدام أربعة عناصر باستخدام مقياس ليكرت من خمس نقاط تم تكييفه من (تشونغ وآخرون، 2022) بما في ذلك عبارات مثل “أشعر بإيجابية تجاه الذكاء الاصطناعي العام”. .
تشير التعزيز الاجتماعي إلى مدى تشجيع التكنولوجيا من قبل الآخرين ضمن مجموعة الإشارة أو الشبكات الاجتماعية الخاصة بهم. تم قياس ذلك باستخدام سبعة عناصر باستخدام مقياس ليكرت من خمس نقاط تم تكييفه من (Compeau & Higgins، 1995) ويتكون من خمسة بيانات مثل تم سؤال المستجيبين عن مدى تشجيع استخدام GenAI من قبل “وظيفتك، مديرك، زملائك، أصدقائك، عائلتك، زملاء العمل، المرؤوسين.” ).
تقييم الثقة هو الحالة النفسية التي تتضمن توقعات إيجابية حول الذكاء الاصطناعي التوليدي (Choung et al., 2022). قمنا بتكييف مقياس من Huh et al. (2005) الذي طلب من المشاركين تقييم الذكاء الاصطناعي التوليدي على أربعة عناصر باستخدام مقياس من خمس نقاط (مقنع، موثوق، جدير بالثقة، وموثوق به). ).
تُقاس الراحة مع التكنولوجيا من خلال قبولها واستخدامها ومستوى الراحة العام معها. هنا، تم قياسها باستخدام مقياس ليكرت المكون من خمس نقاط والمعدل من (روزن وآخرون، 2013) والذي يتكون من ستة عناصر مثل “أشعر أنني أنجز المزيد بفضل التكنولوجيا”. ).
أخيرًا، تضمنت مخاوف الذكاء الاصطناعي تسعة عناصر تم الحصول عليها من تقارير وسائل الإعلام ومناقشات مع العديد من الممارسين. تم تقييم كل عنصر على مقياس من خمسة عناصر وتم الحصول على القياس من خلال متوسط ​​درجة العناصر التسعة (مثل: الانتحال، حقوق الطبع والنشر، المساءلة، الأصالة، إلخ). .

خصائص العينة

بلغ إجمالي حجم العينة 294 من أعضاء هيئة التدريس في التعليم العالي الذين أكملوا الاستبيان. كانت العينة تتكون من أساتذة كاملين ( أستاذ مشارك ( )، أساتذة مساعدين أعضاء هيئة التدريس المقيمون أعضاء هيئة التدريس المساعدين )، وأعضاء هيئة التدريس الزائرين ( ). من العينة، لم يشغلوا منصبًا إداريًا. كان لديهم سنوات متعددة من الخبرة، لـ سنوات من الخبرة، بين 9 و 14 عامًا بين 4 و 8 سنوات بين 15 و 20 سنة. كانت المشاركات إناثاً ( )، الذكور أفضل عدم الإجابة )، وغير ثنائي/آخرون ( ” ). تم تحديد المشاركين على أنهم قوقازيون (69%)، سود أو أمريكيون من أصل أفريقي (7%)، آسيويون أو من جزر المحيط الهادئ (7%)، لاتينيون ( العرق المختلط وآخرون تراوحت أعمارهم بين . فقط من أساتذة الجامعات قد أبلغوا أنهم استخدموا GenAI من قبل.

أسئلة البحث

RQ1: هل يعمل نموذج TAM بشكل مشابه مع GenAI كما هو الحال مع التقنيات الأخرى؟
RQ2: ما هي العلاقة بين نموذج TAM (المنفعة المتصورة والسهولة المتصورة) ونموذج SCT (الكفاءة الذاتية والتعزيز الاجتماعي) في فهم اعتماد الذكاء الاصطناعي التوليدي؟
RQ3: كيف يؤثر الثقة على العلاقة بين نموذج TAM (المنفعة المتوقعة وسهولة الاستخدام المتوقعة) ونموذج SCT (الكفاءة الذاتية والتعزيز الاجتماعي) في فهم اعتماد الذكاء الاصطناعي التوليدي؟

النتائج

تم إجراء سلسلة من التحليلات الإحصائية لاختبار فرضيات الدراسة. تتعلق أولى أسئلة البحث بتقييم الفرضيات المتعلقة بنموذج قبول التكنولوجيا (TAM) مع تكنولوجيا الذكاء الاصطناعي التوليدي (GenAI). بالنسبة للفرضية الأولى، تم إجراء تحليل انحدار متعدد للتنبؤ بمواقف أساتذة الكلية تجاه استخدام GenAI من (أ) الفائدة المدركة (PU) و(ب) السهولة المدركة للاستخدام (PEU). أظهرت النتائج أن كلا المتغيرين توقعا بشكل إحصائي ملحوظ الموقف تجاه الاستخدام. ، محاسبة عن من التباين. الـ PU و PEU تم العثور على أنها تتنبأ بشكل إيجابي بالاتجاه نحو استخدام GenAI. وبالتالي، تم دعم H1 (أ، ب) بالكامل (انظر الجدول 1).
وبالمثل، بالنسبة للفرضية الثانية، تم إجراء تحليل الانحدار المتعدد للتنبؤ بنية أساتذة الجامعات السلوكية لاستخدام GenAI من (أ) الفائدة المتوقعة، (ب) السهولة المتوقعة في الاستخدام، و(ج) المواقف. أظهرت النتائج أن النموذج يفسر من التباين وأن النموذج كان متنبئًا مهمًا للنوايا السلوكية للاستخدام ، . فقط الـ PU والموقف ، “ تم العثور عليها لتتنبأ بشكل إيجابي بالنية السلوكية للاستخدام
الجدول 1 الانحدار المتعدد مع نموذج قبول التكنولوجيا (TAM)
الموقف نية السلوك
توقيع توقيع
الفائدة المدركة 0.655 0.000 0.347 0.000
سهولة الاستخدام المدركة 0.163 0.003 0.645
الموقف 0.462 0.000
GenAI، لكن PEU لم يكن له دلالة كبيرة ( لذا، كانت H2 (أ، ج) مدعومة فقط (انظر الجدول 1).
تنبأت الفرضية الثالثة بمدى استخدام أساتذة الجامعات للذكاء الاصطناعي التوليدي بناءً على سهولة الاستخدام المتوقعة. أظهرت النتائج أن النموذج كان ذا دلالة. ، محاسبة لـ من التباين، و PEU تم العثور على أنه يتنبأ بشكل إيجابي بفائدة الاستخدام المتصور لاستخدام GenAI. وبالتالي، تم دعم H3.
لاختبار فرضية H4، قمنا بإجراء تحليل وساطة باستخدام ماكرو SPSS PROCESS (هايز، 2017) لفحص ما إذا كانت المواقف تتوسط العلاقة بين (أ) الفائدة المتصورة و (ب) سهولة الاستخدام المتصورة على نية سلوك أساتذة الجامعات في استخدام GenAI. باستخدام النموذج 4، قمنا بتشغيل تحليل Bootstrap مع 5000 تكرار افتراضي. بالنسبة للفائدة المتصورة، أشارت النتائج إلى أن النموذج كان متنبئًا مهمًا للموقف [ F(2, 188) مع محاسبة التباين. أظهرت النتائج أن المواقف وسّطت العلاقة بين الاستخدام المفيد ونية السلوك. فترة الثقة من 0.24 إلى 0.56). بينما بالنسبة لـ PEU، أشارت النتائج إلى أن النموذج كان متنبئًا مهمًا للموقف. مع محاسبة التباين. أظهرت النتائج أن المواقف وسّطت العلاقة بين سهولة الاستخدام المتوقعة ونية السلوك. من 0.27 إلى 0.58). وبالتالي، تم دعم H4 (أ، ب) بالكامل فقط (انظر الشكل 2).
السؤال البحثي الثاني وفرضياته المقابلة تهدف إلى دمج وفحص نموذج TAM مع نظرية SCT. الفرضية الخامسة تتعلق بالكفاءة الذاتية. تم إجراء تحليل الانحدار المتعدد المتغيرات لتقييم كيفية تأثير الكفاءة الذاتية على (أ) PEU، (ب) المواقف و(ج) النية السلوكية لاستخدام GenAI. أظهرت النتائج أن الكفاءة الذاتية في GenAI تنبأت بشكل إحصائي ملحوظ بـ PEU. CI [0.090، 0.172])، لكن لا التوجه المتوقع ( )، ولا النية السلوكية لاستخدام GenAI ( وبالتالي، فإن H5a فقط هو المدعوم (انظر الجدول 2).
بالنسبة للفرضية السادسة، تم إجراء تحليل الانحدار المتعدد المتغيرات لفحص تأثير التعزيز الاجتماعي على الاستخدام المتوقع (PU)، والسهولة المتوقعة للاستخدام (PEU)، والموقف، والكفاءة الذاتية. أظهرت النتائج أن التعزيز الاجتماعي توقع بشكل إحصائي ملحوظ جميع المتغيرات؛ الاستخدام المتوقع (PU). سي آي ، بي إي يو سي آي , ) ، الموقف ( ) ، وكفاءة الذات ( ، CI [0.013، 0.702]). وبالتالي، تم دعم جميع H6 (أ، ب، ج، د) بالكامل (انظر الجدول 2).
لاختبار الفرضية السابعة، قمنا بإجراء تحليل وساطة باستخدام ماكرو SPSS PROCESS (هايز، 2017) لفحص ما إذا كانت التعزيز الاجتماعي يتوسط العلاقة بين (أ) سهولة الاستخدام المتوقعة و (ب) الفائدة المتوقعة ونية أساتذة الجامعات السلوكية لاستخدام GenAI. باستخدام النموذج 4، قمنا بإجراء تحليل Bootstrap مع 5000 تكرار افتراضي. بالنسبة لسهولة الاستخدام المتوقعة، أظهرت النتائج أن النموذج كان ذا دلالة. ، وأن التعزيز الاجتماعي وسّط العلاقة بين سهولة الاستخدام المتوقعة و
الشكل 2 تحليل الوساطة TAM (H4)
الجدول 2 مقارنة العوامل المؤثرة على اعتماد التكنولوجيا
الكفاءة الذاتية تعزيز اجتماعي ثقة
ب توقيع توقيع توقيع
الفائدة المدركة 0.551 0.348 0.000 0.525 0.000
سهولة الاستخدام المدركة 0.131 0.000 0.161 0.003 0.316 0.000
الموقف 0.113 0.335 0.000 0.630 0.000
نية السلوك 0.571 0.238 0.003 0.503 0.000
ثقة 0.072 0.011 0.232 0.001
الكفاءة الذاتية 0.357 0.042 0.466 0.011
تعزيز اجتماعي 0.061 0.042 0.254 0.001
الشكل 3 تحليل الوساطة لتعزيز الاجتماعي (H7)
نية السلوك من 0.00 إلى 0.12) (انظر الشكل 3). ومع ذلك، بالنسبة لـ PU، لم يتم إثبات الوساطة لأن معامل الانحدار بين التعزيز الاجتماعي ونية السلوك لم يكن ذا دلالة إحصائية عند ( ) ويسقط الصفر ضمن نطاق فترة الثقة ( -0.07 إلى 0.03 ). يتم تفسير ذلك من خلال التأثير المباشر ( ) تكون شبه مطابقة للتأثير الكلي ( )، مما يشير إلى عدم وجود تأثير غير مباشر. وبالتالي، تم دعم H7a فقط.
السؤال البحثي الثالث وفرضياته المقابلة تتناول إدراك الثقة. بالنسبة للفرضية الثامنة، وجدت تحليل الانحدار المتعدد المتغيرات أن الثقة أثرت بشكل كبير إحصائيًا على الاستخدام المتوقع (PU). ) ، PEU ( سي آي ) ، الموقف ( CI [0.493، ) ، ونية السلوك لاستخدام GenAI ( CI [0.354, 0.652])، كما هو موضح في الجدول 2. وبالتالي، تم دعم H8 (أ، ب، ج، د) بالكامل (انظر الجدول 2).
نظرًا لأن الثقة كانت واحدة من العوامل القوية التي أثرت على نموذج TAM، وللإجابة على الفرضية التاسعة، قمنا بإجراء تحليل تعديل باستخدام النموذج 1 مع ماكرو SPSS PROCESS (هايز، 2017) لتقييم ما إذا كانت الثقة تعدل العلاقة بين الفائدة المتصورة (PU) ونية السلوك تجاه GenAI. بالنسبة للفائدة المتصورة، أشارت النتائج إلى أن النموذج العام كان ذا دلالة إحصائية. . كانت العلاقة بين PU والثقة ذات دلالة )، مما يشير إلى أن العلاقة بين الاستخدام المفيد (PU) ونية السلوك تعتمد على مستوى الثقة. أظهرت التحليلات أن المنحدرات البسيطة عند مستويات مختلفة من الثقة (مثل: مستويات منخفضة، متوسطة، وعالية) جميعها إيجابية، مما يدل على أن الاستخدام المفيد له تأثير إيجابي على نية السلوك عند جميع مستويات الثقة. ومع ذلك، فإن حجم هذا التأثير الإيجابي يتناقص مع زيادة الثقة. وبالتالي، كانت العلاقة بين الاستخدام المفيد ونية السلوك أقوى عندما كانت الثقة منخفضة. ،
الشكل 4 الاعتدال مع الثقة
الشكل 5 الوساطة مع الثقة
مقارنةً عندما كانت مرتفعة . وبالتالي، تم دعم H9 (انظر الشكل 4).
بالنسبة للفرضية العاشرة، قمنا بإجراء تحليل الوساطة باستخدام ماكرو SPSS PROCESS (هايز، 2017) لفحص ما إذا كانت الثقة تتوسط العلاقة بين التعزيز الاجتماعي ونية السلوك لاستخدام GenAI. باستخدام النموذج 4، أظهرت النتائج أن النموذج كان ذا دلالة. وكانت تلك الثقة تتوسط العلاقة بين التعزيز الاجتماعي ونية السلوك من 0.04 إلى 0.19). من الجدير بالذكر أن التأثير غير المباشر فقط هو الذي كان له دلالة إحصائية، بينما التأثير المباشر لم يكن له دلالة إحصائية. لذا، تم دعم H10 (انظر الشكل 5).

نقاش

من خلال دمج نموذج TAM ونظرية SCT، استكشفت هذه الدراسة تصورات أعضاء هيئة التدريس في التعليم العالي ومواقفهم وتجاربهم بشأن الذكاء الاصطناعي التوليدي. كما تفحص الفوائد والعيوب المحتملة، والدوافع والمخاوف، والعوامل وراء اعتماد أو رفض أدوات الذكاء الاصطناعي التوليدي. بشكل عام، تقدم نتائجنا دعماً نظرياً للنموذج المقترح لتقييم قرارات اعتماد الذكاء الاصطناعي التوليدي بين المعالجين في الكلية (انظر الجدول 3).
جدول 3 ملخص فرضيات الدراسة
رقم الفرضية مسارات/علاقات الفرضيات بيتا قيمة P نتيجة
H1-TAM PU → Att 0.655 مدعوم
PEU → انتباه 0.163 مدعوم
H2-TAM 0.347 مدعوم
بيو → بي غير مدعوم
Att → BI 0.462 مدعوم
H3-TAM PEU → PU 0.367 مدعوم
الوساطة H4 أAtt يتوسط
0.39 مدعوم
بيو → بي 0.41 مدعوم
الكفاءة الذاتية (SE) SE → PEU 0.131 مدعوم
SE → Att غير مدعوم
SE → BI غير مدعوم
H6-تعزيز اجتماعي (SR) 0.348 مدعوم
SR → PEU 0.161 مدعوم
SR → انتباه 0.335 مدعوم
SR → SE 0.357 مدعوم
H7- الوساطة SR يتوسط
PEU → BI 0.05 مدعوم
غير مدعوم
H8-ثقة (T) 0.525 مدعوم
T → PEU 0.316 مدعوم
أAtt 0.630 مدعوم
0.503 مدعوم
اعتدال H9 طن
-0.21 مدعوم
H10-الوساطة توسط
0.11 مدعوم

تام

تم العثور على أن جميع الفرضيات المستمدة من نموذج قبول التكنولوجيا (TAM) كانت ذات دلالة إحصائية. أشارت النتائج إلى أن الفائدة المدركة تؤثر بشكل أقوى بكثير على المواقف، وبشكل خاص، على النوايا السلوكية مقارنةً بالسهولة المدركة للاستخدام. وهذا يبرز أهمية إظهار الفوائد العملية لتكنولوجيا GenAI لتشجيع اعتمادها.
من الجدير بالذكر أنه على الرغم من أن PEU لم تظهر تأثيرًا مباشرًا كبيرًا على نوايا أساتذة الجامعات السلوكية لاستخدام GenAI )، أظهر تأثير وساطة غير مباشر كبير على النوايا السلوكية من خلال المواقف. على وجه التحديد، كشفت تحليل الوساطة أن من التأثير الكلي ( )، كان التأثير غير المباشر مسؤولاً عن بينما كان التأثير المباشر نسبيًا ضئيلًا، عند ( ). وبالتالي، يفسر التأثير غير المباشر بشكل رئيسي العلاقات داخل النموذج. تشير هذه النتائج إلى أن تصورات سهولة استخدام GenAI وحدها غير كافية لدفع التبني بين الأساتذة، ولكن عندما تقترن هذه التصورات بمواقف إيجابية تجاه GenAI، يمكن أن تعزز بشكل كبير من احتمالية التبني. وهذا يشير إلى أن الأساتذة قد يحتاجون إلى أسباب ودوافع مقنعة للمشاركة.
مع GenAI. على النقيض من ذلك، تُظهر الفائدة المدركة (PU) تأثيرًا قويًا على النوايا السلوكية لاستخدام GenAI، سواء بشكل مباشر أو غير مباشر من خلال المواقف.

ثقة

بشكل عام، يظهر أن الثقة هي المتنبئ الأكثر أهمية الذي يؤثر على جميع العوامل (PU، PEU، SR، SE، Att) المتعلقة بقرار اعتماد تكنولوجيا GenAI، تليها التعزيز الاجتماعي عن كثب. في هذه الدراسة، تشير الثقة إلى تصورات المصداقية والموثوقية والاعتمادية. تظهر النتائج أن الثقة في التكنولوجيا لها تأثير مباشر على اعتقاد الأساتذة بأنها ستعزز أدائهم، أي تصورات الفائدة. يمكن أن تعزز أيضًا شعورًا بالبساطة، مما يعزز بدوره كفاءتهم الذاتية – ثقتهم في قدرتهم على استخدام التكنولوجيا بفعالية. يمكن أن تؤدي زيادة الثقة إلى مواقف أكثر إيجابية ونوايا أقوى لاعتماد تكنولوجيا GenAI. هذا الأمر ذو صلة خاصة في ضوء المخاوف السائدة المحيطة بهذه التكنولوجيا المثيرة للجدل (غوبتا وآخرون، 2024)؛ يمكن أن يبدد مستوى ثقة الأساتذة مخاوفهم ويشكل بشكل كبير تصوراتهم ومواقفهم وقراراتهم اللاحقة بشأن اعتمادها. وبالتالي، تحتاج المؤسسات التعليمية إلى تقديم معلومات واضحة مع معالجة أي معلومات مضللة لتخفيف الخوف والشك حول GenAI وبناء الثقة.
أظهرت تحليل الاعتدال أن الثقة تؤثر بشكل كبير على العلاقة بين الفائدة المدركة ونية اعتماد تكنولوجيا الذكاء الاصطناعي التوليدي. ومع ذلك، فإن هذا التأثير مشروط، حيث يصبح التأثير أضعف عند مستويات أعلى من الثقة. وهذا يشير إلى أنه عندما يكون لدى الأساتذة مستوى عالٍ من الثقة في التكنولوجيا، قد يعتمدون أقل على فائدتها المدركة أو يفضلون ثقتهم على الفائدة المدركة لتحديد نيتهم في اعتمادها. وعلى العكس، عندما تكون الثقة منخفضة، سيحتاج الأساتذة إلى أسباب أوضح وأكثر إقناعًا للاعتماد، مما يدفعهم للاعتماد بشكل كبير على تصورات الفائدة كعامل تحفيزي لتشجيعهم على اعتماد التكنولوجيا. بشكل عام، يبرز هذا الثقة كعامل قوي يمكن أن يطغى على عوامل أساسية أخرى مثل الفائدة المدركة.

تعزيز اجتماعي

تشير النتائج إلى أن التعزيز الاجتماعي يعد مؤشراً مهماً آخر لمعظم العوامل التي تؤثر على قرارات التبني، خاصة لأن التعزيز الاجتماعي هو عامل شائع عبر نظرية التعلم الاجتماعي (باندورا، 1999) ونظريات تبني التكنولوجيا المختلفة (جرانيك ومارانغونيك، 2019؛ فينكاتيش وآخرون، 2016). هذه النتيجة تتماشى مع تحقيقات مماثلة في مجالات أخرى مثل الرعاية الصحية (روبيلت وآخرون، 2024) والصحافة (ترانغ وآخرون، 2024). وهذا يبرز الدور الحاسم للتفاعلات الاجتماعية والبيئات الاجتماعية في تشكيل تصورات الأساتذة ومواقفهم وقراراتهم بشأن تبني تكنولوجيا الذكاء الاصطناعي التوليدي. على وجه التحديد، وجدت النتائج أن التعزيز الاجتماعي له تأثير مباشر وإيجابي على تصورات الفائدة وسهولة الاستخدام، مما يعزز المواقف الإيجابية ويشجع القرار العام لتبني تكنولوجيا الذكاء الاصطناعي التوليدي. بالإضافة إلى ذلك، يمكن أن يؤثر التعزيز من الدوائر الاجتماعية للفرد بشكل إيجابي على اعتقاد الأساتذة في قدرتهم على استخدام وتبني الذكاء الاصطناعي التوليدي (الكفاءة الذاتية)؛ الاعتقاد بأن “إذا كان الآخرون يستطيعون القيام بذلك، فأنا أيضاً أستطيع” يعمل على تحفيز وتسهيل التبني بين الأقران. لذلك، تشير هذه النتائج إلى أن الديناميات الاجتماعية تلعب دوراً كبيراً في قرارات تبني تكنولوجيا الذكاء الاصطناعي التوليدي. هذا التأثير مهم بالنظر إلى أن تكنولوجيا الذكاء الاصطناعي التوليدي مرتبطة بمخاوف أخلاقية متنوعة، مما يدفع العديد من الأفراد إلى
اقترب منه بحذر. وبالتالي، يمكن أن تؤثر التفاعلات الاجتماعية مع الأقران، والمحادثات، والتعليقات الإيجابية، والتجارب التي يشاركها الآخرون ضمن مجموعاتهم الاجتماعية، والنقاشات على المستويات الإدارية، والكليات، والجامعات جميعها على التصورات، والمواقف، واستخدام التكنولوجيا. وهذا يشير إلى أن التوقعات والسلوكيات لمجموعة اجتماعية يمكن أن تؤثر على ميول الفرد تجاه تكنولوجيا الذكاء الاصطناعي التوليدي وقيمتها المدركة. إذا أصبح استخدام تكنولوجيا معينة شائعًا داخل مجتمع ما، فإن الأفراد يكونون أكثر عرضة لتبنيها.
بالإضافة إلى ذلك، أظهرت تحليل الوساطة أنه مع وجود التعزيز الاجتماعي كوسيط، لم يكن هناك تأثير غير مباشر وساطة بين الفائدة المدركة ونية السلوك، بل كان هناك تأثير مباشر فقط. وهذا يعني أنه بمجرد أن يدرك الأساتذة أن تقنية GenAI مفيدة، سيستخدمونها. لا يحتاجون إلى مزيد من التعزيز من شبكاتهم الاجتماعية. وهذا يفسر لماذا كان التأثير المباشر شبه مطابق للتأثير الكلي. من ناحية أخرى، بالنسبة لسهولة الاستخدام المدركة، كان هناك تأثير وساطة كبير على نية السلوك لاستخدام GenAI. وهذا يشير إلى أن تصورات السهولة تتعزز عندما يتم تعزيزها من قبل الآخرين ضمن الدوائر الاجتماعية الذين استخدموها أو يحملون مواقف إيجابية تجاه التكنولوجيا الجديدة، مما يؤدي إلى زيادة احتمالية التبني.

الثقة والتعزيز الاجتماعي

أظهرت تحليل الوساطة أنه لا يوجد تأثير مباشر من التعزيز الاجتماعي على نية السلوك، ولكن كان هناك تأثير غير مباشر كبير للثقة كوسيط في العلاقة بين التعزيز الاجتماعي ونية السلوك. وهذا يعني أنه بينما لا يؤثر التعزيز الاجتماعي بمفرده بشكل مباشر على النية، فإنه يعزز الثقة، مما يزيد من احتمالية اعتماد التكنولوجيا. وذلك لأن مثل هذه التفاعلات الاجتماعية يمكن أن تكون بمثابة تحقق وتعزيز يعززان ثقة الفرد، كما يمكن أن تساعد في تقليل عدم اليقين أو تخفيف الشكوك، مما يجعل الشخص يشعر بمزيد من الثقة لاستخدامها. ومع ذلك، فإن مجرد تلقي التعزيز الاجتماعي لا يؤدي مباشرة إلى نية استخدام التكنولوجيا لأنه يعتمد على مصداقية المصدر والآراء المسبقة التي يمكن أن تجعلهم متشككين وأكثر مقاومة للآراء الإيجابية التي يشاركها الآخرون.

الكفاءة الذاتية

أشارت النتائج إلى أن الكفاءة الذاتية يمكن أن تؤثر فقط على تصورات سهولة الاستخدام، وهو ما يتماشى مع الأبحاث السابقة التي تظهر علاقة إيجابية بين الكفاءة الذاتية التكنولوجية وسهولة الاستخدام المدركة (هولدن ورادا، 2011). وهذا منطقي لأن المزيد من الثقة في القدرة على استخدام GenAI سيؤدي إلى إحساس بالتمكن وسهولة الاستخدام. ومع ذلك، أشارت النتائج أيضًا إلى أن هذه الثقة وحدها غير كافية لتوليد مواقف إيجابية أو نوايا سلوكية لاستخدام GenAI، مما يشير إلى أن المشكلة قد لا تكمن في القدرة على استخدام التكنولوجيا بل في المخاوف والجدل المستمرين المحيطين باستخدامها. تتماشى هذه النتيجة مع الأبحاث السابقة التي تشير إلى أن العلاقة بين الكفاءة الذاتية والنوايا السلوكية غالبًا ما تكون مختلطة وليست دائمًا ذات دلالة (هولدن ورادا، 2011؛ موتشيغوي وباتاني، 2015). تتشكل المواقف والنوايا السلوكية من مجموعة أوسع من العوامل المعرفية والاجتماعية والسياقية مثل القيم والمعتقدات والدعم المؤسسي والثقة وتأثير الأقران، والتي قد تطغى على تأثير الكفاءة الذاتية وحدها. وفقًا لباندورا (1997)، تؤثر أربعة عوامل على الكفاءة الذاتية بطرق مختلفة: إتقان التجارب، النمذجة الاجتماعية، الإقناع اللفظي، و
الاستجابات العاطفية. تسلط هذه العوامل الضوء على تعقيد الكفاءة الذاتية، حيث تلعب التعزيز الاجتماعي والثقة أدوارًا مهمة في تشكيلها. وهذا يشير إلى أن الكفاءة الذاتية، على الرغم من أهميتها، ليست المحرك الوحيد للمواقف أو النوايا السلوكية، التي تتشكل بدلاً من ذلك من مجموعة أوسع بكثير من الاعتبارات.
تفسير آخر محتمل للمعنى المحدود للكفاءة الذاتية، على الرغم من شعور أعضاء هيئة التدريس بالقدرة على استخدام تكنولوجيا الذكاء الاصطناعي (أي، كفاءة ذاتية عالية)، قد يتساءلون عن مدى ملاءمتها أو فعاليتها. يمكن أن تقوض هذه الشكوك نواياهم لاستخدام التكنولوجيا، مما يساعد على تفسير سبب تأثير الفائدة المدركة بشكل أقوى على المواقف والنوايا السلوكية مقارنةً بسهولة الاستخدام. بالإضافة إلى ذلك، قد تنشأ المقاومة أيضًا من كون أعضاء هيئة التدريس معتادين على طرق التدريس التقليدية، أو المخاوف بشأن زيادة عبء العمل أو الالتزامات الزمنية لاستخدام تكنولوجيا الذكاء الاصطناعي، أو التهديدات المدركة للاستقلالية. لذلك، بينما قد يكون لدى أعضاء هيئة التدريس الكفاءة الذاتية لاستخدام الذكاء الاصطناعي، قد لا يزال لديهم مواقف سلبية أو يفشلون في دمجه في عملهم.

الخاتمة

دمجت هذه الدراسة نموذج TAM ونظرية SCT لاستكشاف تصورات أعضاء هيئة التدريس في التعليم العالي ومواقفهم ودوافعهم ومخاوفهم وتفحص الفوائد المحتملة والعيوب والعوامل وراء اعتماد أو رفض أدوات GenAI. وجدت النتائج دعمًا نظريًا للنموذج المقترح لتقييم اعتماد GenAI. على وجه التحديد، تظهر النتائج أن التواصل بشأن الفوائد العملية لـ GenAI ضروري لاعتماد GenAI. ومع ذلك، فإن سهولة الاستخدام المدركة وحدها غير كافية لتشجيع الاعتماد، ولكن عندما تقترن بمواقف إيجابية والدائرة الاجتماعية، تصبح مؤثرة في تشجيع الاعتماد. علاوة على ذلك، تعتبر الثقة والتعزيز الاجتماعي لاعبين رئيسيين وحاسمين في التأثير على قرارات اعتماد GenAI. هناك ديناميكية اجتماعية متضمنة في عملية اعتماد GenAI يمكن أن تشكل تصورات ومواقف وقرارات الأساتذة. يمكن أن تعزز التفاعلات الاجتماعية الإيجابية الثقة من خلال العمل كتحقق أو تعزيز لوجهات نظر الفرد، وكذلك يمكن أن تقلل من عدم اليقين أو تخفف الشكوك، مما يزيد من احتمالية اعتماد التكنولوجيا.

الآثار والدراسات المستقبلية والقيود

تستكشف هذه الدراسة وجهة نظر أعضاء هيئة التدريس في التعليم العالي بشأن استخدام واعتماد GenAI. يمكن أن يساعد الفهم الأعمق لوجهات نظر المعلمين في تشكيل سياسات التعليم العالي وتوجيه قادة الجامعات وصانعي السياسات والمعلمين في تحسين تجربة التعليم للطلاب الجامعيين في قوة العمل المدفوعة بالذكاء الاصطناعي. نظرًا لأن الثقة تلعب دورًا محوريًا في اعتماد GenAI، تحتاج المؤسسات التعليمية إلى تقديم معلومات شفافة ومعالجة المعلومات المضللة لتقليل الخوف والشكوك المحيطة بالتكنولوجيا، بينما تتواصل بفعالية بشأن فوائدها العملية ووظائفها لبناء المصداقية. نظرًا للدور الكبير للديناميات الاجتماعية في تشكيل مواقف أعضاء هيئة التدريس تجاه GenAI، يمكن أن يؤثر الحوار الاجتماعي البناء والتجارب المشتركة داخل الأقسام والكليات والجامعات بشكل كبير على تصورات التقنيات الجديدة. وبالتالي، يمكن أن يعمق تعزيز بيئة اجتماعية داعمة فهمنا للتأثيرات المحتملة لـ GenAI في البيئات التعليمية، مما يساهم في نهج أكثر اطلاعًا وإنصافًا لتحسين الممارسات التعليمية والنتائج.
يمكن أن تستكشف الدراسات المستقبلية دور الدعم المؤسسي والرسائل المتعلقة باستخدام الذكاء الاصطناعي بشكل أكبر. على الرغم من أن الكفاءة الذاتية كان لها تأثير ضئيل على قرارات اعتماد أعضاء هيئة التدريس، إلا أن التحقيق الإضافي في الأسباب الكامنة ضروري، حيث يمكن أن يؤثر على الاستخدام الفعال لهذه التكنولوجيا ويوفر فهمًا أفضل لسلوكيات أعضاء هيئة التدريس. نظرًا للقضايا الأخلاقية الكبيرة المحيطة بـ GenAI، فإنها تكنولوجيا من المحتمل أن تتبع مسار اعتماد أقل تقليدية، خاصة في التعليم العالي. على سبيل المثال، ستؤدي حالة عدم اليقين بشأن استخدام أعضاء هيئة التدريس للذكاء الاصطناعي في الكتابة التي لا ‘تتجاوز الخط’ إلى جعل الكثيرين يتوقفون. لذلك، من الضروري معالجة المخاوف المتعلقة بالاستخدام المسؤول والأخلاقي لـ GenAI. تواجه المؤسسات أيضًا قرارات مستمرة بشأن توفر أدوات الذكاء الاصطناعي التي غالبًا ما تكون مدمجة في البرمجيات الحالية. بالإضافة إلى ذلك، يوفر استخدام الأساليب النوعية فهمًا أعمق لكيفية إدراك أعضاء هيئة التدريس لتكنولوجيا GenAI، واعتباراتها الأخلاقية، وتأثيرها المحتمل على ممارساتهم التعليمية. يمكن أن يساعد دمج الأسئلة المفتوحة ضمن المجالات المختلفة للاستخدام في توضيح النتائج بشكل أكبر. يجب أن تبحث الأبحاث في كيفية اختلاف اعتماد GenAI عبر التخصصات الأكاديمية والسياقات الثقافية، مع تحديد التحديات والفرص الفريدة داخل كل مجال. نظرًا للتطورات السريعة في GenAI، فإن البحث الطولي ضروري لتقييم كيفية تطور التصورات والاستخدامات لـ GenAI بمرور الوقت.
تحد من استراتيجية العينة المستهدفة في المجالات غير STEM عمومًا قابلية تعميم هذه الدراسة. لا شك أن المعلمين خارج نطاق هذه الفئة سيختلفون في آرائهم حول دور GenAI في عملهم. علاوة على ذلك، تم إجراء الدراسة مع جامعتين حكوميتين متوسطة الحجم تركزان على البحث في الولايات المتحدة، مما قد يحد من قابلية تعميمها على سياقات تعليمية أخرى مقارنة بالمؤسسات الأصغر، والجامعات الخاصة، والمراكز التعليمية المكثفة، والإعدادات الدولية، مما قد يؤدي إلى نتائج مختلفة. قد تؤدي الاعتماد على بيانات التقرير الذاتي بشأن مواضيع قد تكون مثيرة للجدل إلى استجابات قد تُعتبر أكثر ملاءمة على حساب الدقة. بالإضافة إلى ذلك، قد لا تلتقط دراسة التركيز على نقطة زمنية معينة المواقف المتطورة أو التغيرات في السلوكيات، خاصة مع تقدم تكنولوجيا GenAI بسرعة. قد يكون قيد آخر محتمل هو أن تحيز الاختيار الذاتي يمكن أن يؤثر على النتائج ويحد من قابلية تعميم النتائج. وبالتالي، قد لا تعكس النتائج بشكل كامل تنوع وجهات النظر أو السلوكيات أو التجارب داخل الفئة السكانية الأوسع. أخيرًا، بينما درسنا دور الثقة والتعزيز الاجتماعي، يمكن أن تؤثر عدة عوامل على اعتماد الذكاء الاصطناعي من قبل أعضاء هيئة التدريس مثل الدعم المؤسسي والثقافة، والتدريب والتطوير المهني، والتخصص أو مجالات التخصص.

الملحق

انظر الجداول 4 و 5 و 6 و 7 و 8 و 9 و 10
الجدول 4 قائمة الاختصارات
الاختصار المصطلح
GenAI الذكاء الاصطناعي التوليدي
PU الفائدة المدركة
PEU سهولة الاستخدام المدركة
TAM نموذج قبول التكنولوجيا
UTAUT النظرية الموحدة لقبول واستخدام التكنولوجيا
SCT نظرية التعلم الاجتماعي
BI النوايا السلوكية
Att الموقف
SR التعزيز الاجتماعي
SE الكفاءة الذاتية
الجدول 5 TAM والموقف
النموذج المعاملات غير القياسية المعاملات القياسية t Sig
B خطأ معياري بيتا
(ثابت) 0.226 0.268 0.844 0.400
PU 0.749 0.061 0.655 12.221 0.000
PEU 0.235 0.077 0.163 3.040 0.003
أ. المتغير التابع: الموقف
الجدول 6 نموذج قبول التكنولوجيا والنوايا السلوكية
نموذج المعاملات غير المعيارية معاملات بيتا الموحدة ت توقيع
ب خطأ معياري
(ثابت) 1.139 0.258 ٤.٤٢٤ 0.000
بيو 0.035 0.076 0.024 0.462 0.645
PU 0.399 0.079 0.347 ٥.٠٦٥ 0.000
الموقف 0.464 0.070 0.462 6.633 0.000
أ. المتغير التابع: نية السلوك
الجدول 7 PU و PEU
نموذج المعاملات غير المعيارية المعاملات المعيارية ت توقيع
ب خطأ معياري بيتا
1 (ثابت) 1.613 0.298 ٥.٤١٤ 0.000
بيو 0.471 0.086 0.367 ٥.٤٨٧ 0.000
الجدول 8 الكفاءة الذاتية
المتغير التابع معامل ب خطأ معياري ت توقيع فترة الثقة 95%
حد أدنى حد أعلى
بي يو اعتراض 3.116 0.199 15.639 0.000 ٢.٧٢٣ ٣.٥٠٩
الكفاءة الذاتية 0.017 0.029 0.597 0.551 -0.040 0.074
بيو اعتراض ٢.٥٣٨ 0.143 17.710 0.000 ٢.٢٥٥ 2.820
الكفاءة الذاتية 0.131 0.021 6.334 0.000 0.090 0.172
أATT اعتراض ٣.١٠٢ 0.226 ١٣٫٧٠٠ 0.000 2.655 ٣.٥٤٨
الكفاءة الذاتية 0.052 0.033 1.593 0.113 -0.012 0.117
ذكاء الأعمال اعتراض ٤.٠٢٠ 0.229 17.569 0.000 ٣.٥٦٩ ٤.٤٧١
الكفاءة الذاتية 0.019 0.033 0.568 0.571 -0.046 0.084
ريال سعودي اعتراض ٢.٢٢١ 0.205 10.816 0.000 1.816 ٢.٦٢٦
الكفاءة الذاتية 0.061 0.030 2.046 0.042 0.002 0.119
ثقة اعتراض 2.330 0.195 11.944 0.000 1.945 ٢.٧١٥
الكفاءة الذاتية 0.072 0.028 ٢.٥٦٤ 0.011 0.017 0.128
أ. تم حسابه باستخدام ألفا
الجدول 9 التعزيز الاجتماعي
المتغير التابع معامل ب خطأ معياري ت توقيع فترة الثقة 95%
حد أدنى حد أعلى
PU اعتراض ٢.٣١٧ 0.181 12.797 0.000 1.960 ٢.٦٧٤
تعزيز اجتماعي 0.348 0.065 ٥.٣٤٥ 0.000 0.220 0.477
بيو اعتراض ٢.٩٧٣ 0.150 19.797 0.000 2.677 ٣.٢٦٩
تعزيز اجتماعي 0.161 0.054 ٢.٩٨٤ 0.003 0.055 0.268
أATT اعتراض ٢.٥٦٦ 0.211 12.136 0.000 ٢.١٤٩ ٢.٩٨٣
تعزيز اجتماعي 0.335 0.076 ٤.٤٠٠ 0.000 0.185 0.485
ذكاء الأعمال اعتراض ٣.٥٢٠ 0.218 ١٦٫١٥٨ 0.000 ٣.٠٩٠ ٣.٩٤٩
تعزيز اجتماعي 0.238 0.078 ٣.٠٣٧ 0.003 0.083 0.393
ثقة اعتراض ٢.١٩٤ 0.187 11.701 0.000 1.824 ٢.٥٦٤
تعزيز اجتماعي 0.232 0.067 ٣.٤٤٧ 0.001 0.099 0.366
SE اعتراض ٥.٦١٠ 0.486 11.553 0.000 ٤.٦٥٢ 6.568
تعزيز اجتماعي 0.357 0.175 2.046 0.042 0.013 0.702
جدول 10 الثقة
المتغير التابع معامل ب خطأ معياري ت توقيع فترة الثقة 95%
حد أدنى حد أعلى
بي يو اعتراض 1.756 0.183 9.580 0.000 1.395 2.118
ثقة 0.525 0.062 8.428 0.000 0.402 0.648
بيو اعتراض ٢.٥٠٩ 0.156 ١٦٫٠٧٤ 0.000 2.201 2.817
ثقة 0.316 0.053 ٥.٩٦١ 0.000 0.212 0.421
أATT اعتراض 1.676 0.205 8.158 0.000 1.271 ٢.٠٨١
ثقة 0.630 0.070 9.024 0.000 0.493 0.768
ذكاء الأعمال اعتراض ٢.٧٣٢ 0.222 12.294 0.000 2.294 ٣.١٧٠
ثقة 0.503 0.076 6.664 0.000 0.354 0.652
ريال سعودي اعتراض 1.905 0.217 8.773 0.000 1.477 ٢.٣٣٣
ثقة 0.254 0.074 ٣.٤٤٧ 0.001 0.109 0.400
SE اعتراض ٥.٢٤١ 0.534 9.810 0.000 ٤.١٨٧ ٦.٢٩٥
ثقة 0.466 0.182 ٢.٥٦٤ 0.011 0.107 0.824
أ. تم حسابه باستخدام ألفا

شكر وتقدير

غير قابل للتطبيق.

مساهمات المؤلفين

AS: تصور، تصميم العمل، الحصول على البيانات، تحليل البيانات، تفسير البيانات، صياغة العمل، الموافقة على النسخة المقدمة (وأي نسخة معدلة بشكل كبير تتضمن مساهمة المؤلف في الدراسة)؛ وافق على أن يكون مسؤولاً شخصياً عن مساهماته الخاصة وأن يضمن أن الأسئلة المتعلقة بدقة أو نزاهة أي جزء من العمل، حتى تلك التي لم يكن المؤلف متورطاً فيها شخصياً، يتم التحقيق فيها وحلها بشكل مناسب، وتوثيق الحل في الأدبيات. KH: تصميم العمل، الحصول على البيانات، تفسير البيانات، صياغة العمل، الموافقة على النسخة المقدمة (وأي نسخة معدلة بشكل كبير تتضمن مساهمة المؤلف في الدراسة)؛ وافق على أن يكون مسؤولاً شخصياً عن مساهماته الخاصة وأن يضمن أن الأسئلة المتعلقة بدقة أو نزاهة أي جزء من العمل، حتى تلك التي لم يكن المؤلف متورطاً فيها شخصياً، يتم التحقيق فيها وحلها بشكل مناسب، وتوثيق الحل في الأدبيات.

تمويل

لم يتم دعم البحث من قبل منظمات التمويل الخارجية.

توفر البيانات والمواد

تتوفر مجموعات البيانات المستخدمة و/أو التي تم تحليلها خلال الدراسة الحالية من المؤلف المراسل عند الطلب المعقول.

الإعلانات

لم يتم نشر هذه المخطوطة أو تقديمها للمراجعة إلى أي جهة أخرى. تم مراجعة الدراسة والموافقة عليها من قبل مكتب نزاهة البحث بجامعة نيفادا، لاس فيغاس.

المصالح المتنافسة

يعلن المؤلفون أنهم ليس لديهم مصالح متنافسة مالية أو غير مالية.
تاريخ الاستلام: 25 أكتوبر 2024 تاريخ القبول: 10 فبراير 2025
نُشر على الإنترنت: 14 مارس 2025

References

Bandura, A. (1997). Self-efficacy: The exercise of control. Freeman.
Bandura, A. (1999). Social cognitive theory: An agentic perspective. Asian Journal of Social Psychology, 2(1), 21-41. https:// doi.org/10.1111/1467-839X. 00024
Bandura, A. (2002). Social cognitive theory in cultural context. Applied Psychology, 51(2), 269-290. https://doi.org/10. 1111/1464-0597.00092
Barrett, A., & Pack, A. (2023). Not quite eye to A.I.: Student and teacher perspectives on the use of generative artificial intelligence in the writing process. International Journal of Educational Technology in Higher Education, 20(1), 59. https://doi.org/10.1186/s41239-023-00427-0
Bower, M., Torrington, J., Lai, J. W. M., Petocz, P., & Alfano, M. (2024). How should we change teaching and assessment in response to increasingly powerful generative Artificial Intelligence? Outcomes of the ChatGPT teacher survey. Education and Information Technologies. https://doi.org/10.1007/s10639-023-12405-0
Buskens, V. (2020). Spreading information and developing trust in social networks to accelerate diffusion of innovations. Trends in Food Science Technology, 106, 485-488. https://doi.org/10.1016/j.tifs.2020.10.040
Chiu, T. K. F. (2024). Future research recommendations for transforming higher education with generative AI. Computers and Education: Artificial Intelligence, 6, 100197. https://doi.org/10.1016/j.caeai.2023.100197
Choung, H., David, P., & Ross, A. (2022). Trust in AI and its role in the acceptance of AI technologies. International Journal of Human-Computer Interaction, 39(9), 1727-1739. https://doi.org/10.1080/10447318.2022.2050543
Compeau, D. R., & Higgins, C. A. (1995). Application of social cognitive theory to training for computer skills. Information Systems Research, 6(2), 118-143. https://doi.org/10.1287/isre.6.2.118
Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. M/S Quarterly, 13(3), 319-340. https://doi.org/10.2307/249008
Davis, F. D., Bagozzi, R. P., & Warshaw, P. R. (1989). User acceptance of computer technology: A comparison of two theoretical models. Management Science, 35(8), 982-1003. https://doi.org/10.1287/mnsc.35.8.982
Davis, F. D., Granić, A., & Marangunić, N. (2024). The technology acceptance model: 30 years of TAM. Springer International Publishing AG.
Granić, A., & Marangunić, N. (2019). Technology acceptance model in educational context: A systematic literature review. British Journal of Educational Technology, 50(5), 2572-2593. https://doi.org/10.1111/bjet. 12864
Gupta, R., Nair, K., Mishra, M., Ibrahim, B., & Bhardwaj, S. (2024). Adoption and impacts of generative artificial intelligence: Theoretical underpinnings and research agenda. International Journal of Information Management Data Insights, 4(1), 100232. https://doi.org/10.1016/j.jjimei.2024.100232
Hartley, K., Hayak, M., & Ko, U. H. (2024). Artificial intelligence supporting independent student learning: An evaluative case study of ChatGPT and learning to code. Education Sciences, 14(2), 120.
Hayes, A. F. (2017). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford publications.
Holden, H., & Rada, R. (2011). Understanding the influence of perceived usability and technology self-efficacy on teachers’ technology acceptance. Journal of Research on Technology in Education, 43(4), 343-367. https://doi.org/ 10.1080/15391523.2011.10782576
Huh, J., DeLorme, D. E., & Reid, L. N. (2005). Factors affecting trust in on-line prescription drug information and impact of trust on behavior following exposure to DTC advertising. Journal of Health Communication, 10(8), 711-731.
Labadze, L., Grigolia, M., & Machaidze, L. (2023). Role of AI chatbots in education: Systematic literature review. International Journal of Educational Technology in Higher Education, 20(1), 56. https://doi.org/10.1186/ s41239-023-00426-1
Lim, W. M., Gunasekara, A., Pallant, J. L., Pallant, J. I., & Pechenkina, E. (2023). Generative AI and the future of education: Ragnarök or reformation? A paradoxical perspective from management educators. The International Journal of Management Education, 21(2), 100790. https://doi.org/10.1016/j.jime.2023.100790
Marikyan, D., Papagiannidis, S., & Stewart, G. (2023). Technology acceptance research: Meta-analysis. Journal of Information Science. https://doi.org/10.1177/01655515231191177
McClain, C. (2024). Americans’ use of ChatGPT is ticking up, but few trust its election information. Pew Research Center. https://www.pewresearch.org/short-reads/2024/03/26/americans-use-of-chatgpt-is-ticking-up-but-few-trust-its-election-information/
Mollick, E. (2024). Co-Intelligence: Living and Working with AI. Penguin Publishing Group. http://ebookcentral.proquest. com/lib/unlv/detail.action?docID=30848006
Motshegwe, M., & Batane, T. (2015). Factors influencing instructors’ attitudes toward technology integration. Journal of Educational Technology Development and Exchange (JETDE). https://doi.org/10.18785/jetde.0801.01
Pan, X. (2020). Technology acceptance, technological self-efficacy, and attitude toward technology-based self-directed learning: Learning motivation as a mediator. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2020.564294
Roppelt, J. S., Kanbach, D. K., & Kraus, S. (2024). Artificial intelligence in healthcare institutions: A systematic literature review on influencing factors. Technology in Society, 76, 102443. https://doi.org/10.1016/j.techsoc.2023.102443
Rosen, L. D., Whaling, K., Carrier, L. M., Cheever, N. A., & Rokkum, J. (2013). The media and technology usage and attitudes scale: An empirical investigation. Computers in Human Behavior, 29(6), 2501-2511.
Shieh, A., Tran, B., He, G., Kumar, M., Freed, J. A., & Majety, P. (2024). Assessing ChatGPT 4.0’s test performance and clinical diagnostic accuracy on USMLE STEP 2 CK and clinical case reports. Scientific Reports, 14(1), 9330. https://doi.org/10. 1038/s41598-024-58760-x
Steiss, J., Tate, T., Graham, S., Cruz, J., Hebert, M., Wang, J., Moon, Y., Tseng, W., Warschauer, M., & Olson, C. (2024). Comparing the quality of human and ChatGPT feedback of students’ writing. Learning and Instruction, 91, 101894. https:// doi.org/10.1016/j.learninstruc.2024.101894
Tanantong, T., & Wongras, P. (2024). A UTAUT-based framework for analyzing users’ intention to adopt artificial intelligence in human resource recruitment: A case study of Thailand. Systems, 12(1), 28.
Trang, T. T. N., Chien Thang, P., Hai, L. D., Phuong, V. T., & Quy, T. Q. (2024). Understanding the adoption of artificial intelligence in journalism: An empirical study in vietnam. SAGE Open, 14(2), 21582440241255240. https://doi.org/10.1177/ 21582440241255241
Venkatesh, V., Thong, J. Y. L., & Xu, X. (2016). Unified Theory of Acceptance and Use of Technology: A Synthesis and the Road Ahead (SSRN Scholarly Paper 2800121). https://papers.ssrn.com/abstract=2800121
Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User acceptance of information technology: Toward a unified view. MIS Quarterly, 27(3), 425-478. https://doi.org/10.2307/30036540
Wu, K., Zhao, Y., Zhu, Q., Tan, X., & Zheng, H. (2011). A meta-analysis of the impact of trust on technology acceptance model: Investigation of moderating influence of subject and context type. International Journal of Information Management, 31(6), 572-581. https://doi.org/10.1016/j.jijnfomgt.2011.03.004
Youk, S., & Park, H. S. (2023). Who is (communicatively more) responsible behind the wheel? Applying the theory of communicative responsibility to TAM in the context of using navigation technology. Human-Machine Communication, 6(1), 11.
Yusuf, A., Pervin, N., & Román-González, M. (2024). Generative AI and the future of higher education: A threat to academic integrity or reformation? Evidence from multicultural perspectives. International Journal of Educational Technology in Higher Education, 21(1), 21. https://doi.org/10.1186/s41239-024-00453-6

ملاحظة الناشر

تظل شركة سبرينجر ناتشر محايدة فيما يتعلق بالمطالبات القضائية في الخرائط المنشورة والانتماءات المؤسسية.

  1. © المؤلفون 2025. الوصول المفتوح. هذه المقالة مرخصة بموجب رخصة المشاع الإبداعي للاستخدام والنسب 4.0 الدولية، التي تسمح بالاستخدام والمشاركة والتكيف والتوزيع وإعادة الإنتاج بأي وسيلة أو صيغة، طالما أنك تعطي الائتمان المناسب للمؤلفين الأصليين والمصدر، وتوفر رابطًا لرخصة المشاع الإبداعي، وتوضح ما إذا كانت هناك تغييرات قد أُجريت. الصور أو المواد الأخرى من طرف ثالث في هذه المقالة مشمولة في رخصة المشاع الإبداعي الخاصة بالمقالة، ما لم يُشار إلى خلاف ذلك في سطر الائتمان للمواد. إذا لم تكن المادة مشمولة في رخصة المشاع الإبداعي الخاصة بالمقالة وكان استخدامك المقصود غير مسموح به بموجب اللوائح القانونية أو يتجاوز الاستخدام المسموح به، فستحتاج إلى الحصول على إذن مباشرة من صاحب حقوق الطبع والنشر. لعرض نسخة من هذه الرخصة، قم بزيارة http:// creativecommons.org/licenses/by/4.0/.
  2. أ. المتغير التابع: PU

Journal: International Journal of Educational Technology in Higher Education, Volume: 22, Issue: 1
DOI: https://doi.org/10.1186/s41239-025-00511-7
Publication Date: 2025-03-13

Artificial intelligence and communication technologies in academia: faculty perceptions and the adoption of generative AI

Aya Shata and Kendall Hartley²(D)

*Correspondence:
aya.shata@unlv.edu
Hank Greenspun School of Journalism and Media Studies, University of Nevada, Las Vegas, 4505 S. Maryland Parkway, Box #45007, Las Vegas, NV 89154-5007, USA
Department of Teaching and Learning, University of Nevada, Las Vegas, 4505 S. Maryland Parkway, Box #453005, Las Vegas, NV 89154-3005, USA

Abstract

Artificial intelligence (AI) is ushering in an era of potential transformation in various fields, especially in educational communication technologies, with tools like ChatGPT and other generative AI (GenAI) applications. This rapid proliferation and adoption of GenAI tools have sparked significant interest and concern among college professors, who are dealing with evolving dynamics in digital communication within the classroom. Yet, the effect and implications of GenAI in education remain understudied. Therefore, this study employs the Technology Acceptance Model (TAM) and the Social Cognitive Theory (SCT) as theoretical frameworks to explore higher education faculty’s perceptions, attitudes, usage, and motivations, as the underlying factors that influence their adoption or rejection of GenAI tools. A survey was conducted among fulltime higher education faculty members ( ) recruited from two mid-size public universities in the US. Results found that college professors’ perceived usefulness of AI predicted their attitudes and intention to use and adopt the technology, more than their perceived ease of use. Trust and social reinforcement strongly influenced college professors’ GenAI adoption decisions and acted as significant mediators to better understand the relationship between TAM and SCT. Findings emphasized the power of social dynamics in shaping professors’ self-efficacy, attitudes, and use of GenAI. Trust enhances peer influence and affects how perceived usefulness shapes users’ willingness to adopt technology, whereas self-efficacy has a minimal impact. This research provides valuable insights that inform higher education policies aimed at improving the educational experience for college students in an AI-driven workforce.

Keywords: Artificial intelligence, Communication technology, Higher education, Technology adoption, Faculty, Trust, Self-efficacy, Social influence

Introduction

Generative AI (GenAI) has quickly become particularly consequential in higher education. Recent research indicates that of adults aged have used ChatGPT (McClain, 2024). This is up from just nine months earlier. Higher education faculty face a rapidly changing classroom dynamic as these widely used, multimodal tools raise numerous opportunities and challenges. This study investigates these opportunities and challenges by exploring the current use and perspectives of higher education
faculty regarding GenAI, the most recent additions to the educational communications and media technologies toolbox.
Generative AI (GenAI) is defined as a technology that leverages deep learning models to generate human-like content (e.g., images, words) in response to prompts (Lim et al., 2023). Examples of interactive text-based GenAI include OpenAI’s ChatGPT, Google’s Gemini, Anthropic’s Claude, and Meta.AI. Multimodal GenAI tools are also becoming more widespread and are changing the landscape of educational media creation. Image development tools like DALL-E or Midjourney can be used to create visual content. Video development is rapidly advancing with tools such as Open AI’s Sora and RunwayML. Taken together, the potential for creating more engaging and diverse digital educational materials is substantial.
Like most arenas, applications in the higher education context are rapidly emerging, and ever-increasing capabilities make it difficult to identify the most salient use cases. Teaching examples include the development of case studies and assessment items for a given topic. Research examples include support for qualitative data analysis. Logistical support includes the capacity to generate captions for images and videos. The latest LLM models have also introduced an incredible leap in the accuracy of speech-to-text transcription.
While artificial intelligence is not new and has been studied for years (decades in some fields), the rapid proliferation and adoption of GenAI tools have generated widespread concern and interest among educators and researchers (Yusuf et al., 2024). This interest has moved AI implications beyond the original academic disciplines such as computer science, information management, and other STEM fields. The implications for communication and education technologies beyond the academy’s STEM disciplines are now garnering greater attention and are the target of this study. Our focus is on the implications for non-STEM fields where academicians likely do not have a deep grasp of the underlying technology. While this limits the generalizability, it provides valuable insights into the emerging impact of artificial intelligence on disciplines that have traditionally been less influenced by such technologies.

Literature review

Al in higher education

Artificial Intelligence is not a new phenomenon; its applications in higher education have been studied for decades (Hartley et al., 2024). Recent advances in the technology underlying AI applications have dramatically improved general capabilities (Mollick, 2024). These advances have introduced numerous new challenges and opportunities for higher education faculty (Yusuf et al., 2024).
The role of GenAI in academia and its implications for faculty are emerging and complex. One researcher conceptualizes the use of GenAI as ‘co-intelligence’ (Mollick, 2024), which is consistent with the admonition that these new tools are best conceived as assistants or collaborators. Researchers have also identified its paradoxes, as it can be viewed as supportive and adversarial, effective and inaccurate, popular and banned (Lim et al., 2023). Academic tasks such as teaching, service, and research all intersect with GenAI. Early research utilizing GenAI to accomplish tasks such as providing writing feedback has fallen short compared to human evaluators (Steiss et al., 2024). However, the tools
are improving rapidly as is our understanding of how and when to utilize ChatGPT. The study noted above utilized ChatGPT version 3.5. A study comparing ChatGPT version 3.5 with 4.0 on the United States Medical Licensing Exam found marked improvement. 3.5 demonstrated an accuracy of , while 4.0 leaped to (Shieh et al., 2024). Whether or not this improvement translates into higher scores for students remains unclear.

Faculty and student perceptions of AI

Some have argued that the responsibility of preparing students for the future dictates preparation for a “society powered by AI” (Chiu, 2024). A rapidly expanding body of research has provided initial insights into GenAI’s import to higher education. In a comparison of faculty and student perceptions of writing with AI tools, researchers found disagreements regarding the acceptable use of GenAI (Barrett & Pack, 2023). Interestingly, these differences indicated a more positive view of GenAI by teachers than students. Work prior to the introduction of ChatGPT indicates that positive views of GenAI for learning might be justified. In a meta-analysis of the use of text-based chatbots in education, researchers found numerous advantages for students (Labadze et al., 2023). More recent work has supported the finding of improved performance in the short-term, however, there are also reasons to believe that GenAI use is better viewed as a ‘crutch’ that, when removed, can result in decreased understanding in the longer term (Lim et al., 2023). A common concern among educators is the potential for Gen AI to result in a decline in student’s cognitive and logical skills (Yusuf et al., 2024).

Technology adoption

To understand the interaction between higher education faculty and GenAI, it is useful to look at technology adoption frameworks. The Technology Acceptance Model (TAM) suggests that technology acceptance is strongly influenced by its perceived ease of use (PEU) and perceived usefulness (PU) (Davis, 1989; Granić & Marangunić, 2019; Venkatesh et al., 2003). Faculty perceptions of ease of use and usefulness are thus essential avenues to explore GenAI’s implications for higher education. Another relevant framework is the Unified Theory of Acceptance and Use of Technology (UTAUT), which acknowledges the value of TAM constructs, and adds social influence and facilitating conditions to drive behavioral intention to use a technology (Venkatesh et al., 2016). To explore the social and personal factors, the Social Cognitive Theory (SCT) serves as another theoretical framework that explains the technology acceptance process (Bandura, 1999). In this study, we combine TAM and SCT to have a better understanding of college professors’ uses of GenAI.
In brief, the behavioral intention to use technology will be directly related to the attitude towards the technology, which is in turn influenced by the perceived ease of use (PEU) and the perceived usefulness (PU) of the technology. In addition, other factors discussed below will contribute to the adoption equation.

Attitude

The user’s attitude towards the technology and innovation can significantly influence the adoption of new technologies. Users’ attitudes towards technological advancements
are crucial in determining whether or not they will embrace innovations (Davis, 1989; Granić & Marangunić, 2019). Regarding AI, individuals’ attitudes and comfort levels can greatly influence acceptance and integration into various aspects of the education system (Bower et al., 2024). Preliminary studies suggest that faculty attitudes toward Gen AI are mixed (Bower et al., 2024). While a majority of faculty perceive the potential for major impact, these same faculty are more familiar with the technology.

PEU/PU

One’s attitude towards technology can be influenced by its perceived ease of use and its perceived usefulness (Davis et al., 1989; Granić & Marangunić, 2019). These factors become critical as higher education faculty consider the adoption and inclusion of communication technologies.
The arrival of ChatGPT in November 2022 was significant in part because it introduced an approachable avenue for people to interact with an advanced large language model. Prior to this, Generative AI application users were largely limited to those with coding knowledge or domain-specific expertise. However, the perceived ease of use is somewhat limited in that the early interfaces were predominantly text-based. As the interfaces have developed to better match user needs, perceptions have and will improve. For example, ChatGPT, Claude and Gemini can now display preview windows that include working drafts of images, charts, tables, or code, substantially improving the ease of use (Gupta et al., 2024).

Self-efficacy

When considering attitudes and perceptions of new technology, a helpful framework is Bandura’s social cognitive theory, which postulates the importance of content-specific self-efficacy in pursuing knowledge (Bandura, 2002; Venkatesh et al., 2003). Self-efficacy refers to an individual’s belief in their ability to succeed in a given context or content area. Technology self-efficacy can play an essential mediating role in technology acceptance (Granić & Marangunić, 2019; Holden & Rada, 2011). For example, perceived ease of use was associated with self-efficacy in undergraduate students engaged in technol-ogy-supported self-directed learning (Pan, 2020). The intention to use and adopt GenAI tools is likely similarly influenced by self-efficacy for these emerging technologies. Users who have committed significant time to its use will be more likely to confidently use GenAI compared to those who have had limited engagement.

Social reinforcement

Research has also addressed the important role of social reinforcement. This can be defined as the degree to which users believe that their social networks feel it is important that they adopt a particular technology (Marikyan et al., 2023). While a recent meta-analysis concluded that the social influence association with behavior was nonsignificant, it did find an overall significant impact on attitudes (Marikyan et al., 2023). The rapid adoption of GenAI, major investments by major corporations, and extensive media coverage have resulted in substantial direct and indirect pressure on individuals toward adoption. These influences, of course, vary between cultures and disciplines. For example, the impact of GenAI on college composition courses has garnered substantial
negative attention. Social cognitive theory acknowledges our enhanced capacity for observational learning and the vital role of social interactions (Bandura, 1999). This capacity, combined with the widespread discussion, will likely influence users’ perceptions, attitudes, and self-efficacy toward using GenAI. In line with prior work in technology adoption (Venkatesh et al., 2016), we operationalize social influence as social reinforcement, reflecting the impact of others’ opinions and behaviors on an individual’s decision to use the technology.

Trust

Trust becomes a key consideration when investigating the adoption of new technologies. Trust can be described as the willingness of an individual to accept information from another, with the expectation that the information provider is accurate (Wu et al., 2011). A meta-analysis of technology adoption studies identified trust as strongly associated with PEU and PU (Marikyan et al., 2023). For example, Choung et al. (2022) explored the role of trust in adopting AI voice assistant technology, and found trust significantly influenced both perceived usefulness of the AI voice assistant and users’ attitudes toward it, which subsequently elevated usage intention. Similarly, Tanantong and Wongras (2024) examined the use of AI in recruitment by human resources and found that while trust does not affect intention directly, it significantly influenced both perceived usefulness and ease of use. These findings underscore that trust has a significant impact on perceived ease of use, perceived usefulness, and attitudes toward GenAI.
Social networks and trust are integral for the diffusion of new technologies (Busken, 2020). According to Busken, once individuals recognize the potential benefits of a technology, it is necessary to build trust that the innovation will deliver the promised advantages. To achieve this, continuous reinforcement from social circles are key for developing trust in technology. This suggests that social reinforcement can influence trust perceptions which in turn influence the likelihood of technology adoption, particularly in the context of GenAI. Overall, trust becomes uniquely important when it comes to GenAI adoption, given the widespread concerns about the propensity towards hallucinations in the major models.

Current study

The purpose of this study is to improve our understanding of higher education faculty’s perceptions, usage, and attitudes toward GenAI, with a focus on identifying motivations and barriers through the framework of existing research and technology adoption theories. In particular, we investigate the technological, social and psychological factors that have been demonstrated as relevant. These factors include those identified in TAM and SCT in addition to perceptions of Trust (Fig. 1).

Proposed research questions and hypotheses

RQ1: Does the TAM model perform similarly with GenAl as it does with other technologies?
H1 (a) Perceived Usefulness and (b) Perceived Ease of Use will positively influence College Professors’ attitudes towards GenAI.
Fig. 1 Conceptual Framework and Hypotheses Summary
H2 (a) Perceived Usefulness, (b) Perceived Ease of Use, and (c) College Professors’ attitudes will positively influence behavioral intention to use GenAI.
H3 Perceived Ease of Use will positively influence College Professors’ Perceived Usefulness of using GenAI.
H4 College Professors’ attitudes toward GenAI will mediate the relationship between (a) Perceived Usefulness and (b) Perceived Ease of Use on their Behavioral Intention to Use GenAI.
RQ2: What is the relationship between the TAM model (PU and PEU) and the SCT (self-efficacy and social reinforcement) in understanding GenAI adoption?
H5 GenAI self-efficacy positively influences (a) Perceived Ease of Use, (b) College Professors’ attitudes, and (c) behavioral intention to use GenAI.
H6 Social Reinforcement positively influences (a) Perceived Usefulness, (b) Perceived Ease of Use, (c) College Professors’ attitudes, and (d) self-efficacy toward GenAI.
H7 Social reinforcement will mediate the relationship between (a) Perceived Ease of Use, and (b) Perceived Usefulness on college professors’ behavioral intentions to use GenAI.
RQ3: How does Trust influence the relationship between the TAM model (PU and PEU) and the SCT (self-efficacy and social reinforcement) in understanding GenAl adoption?
H8 Trust in GenAI positively influences (a) Perceived Usefulness, (b) Perceived Ease of Use, (c) attitudes, and (d) behavioral intention toward GenAI.
H9 Trust will moderate the relationship between Perceived Usefulness and college professors’ behavioral intentions to use GenAI.
H10 Trust will mediate the relationship between social reinforcement and college professors’ behavioral intentions to use GenAI.

Methods

This study employed a survey design to explore higher education faculty’s perceptions, attitudes, uses, and experiences regarding GenAI. The primary objective was to examine the potential motivations, barriers and factors underlying the adoption or rejection of GenAI tools among university professors.

Participants

An online survey was designed using Qualtrics with full-time higher education faculty members recruited from two mid-size public U.S. universities, one on the East Coast and one in the Southwest, to participate in this study. The sample represented all faculty members in the social sciences and humanities disciplines as they use GenAI in a similar manner that differs fundamentally from its application in STEM disciplines.

Procedures

The study procedures started by creating an email list for all faculty members obtained from the email directory available online, then filtered by discipline to include only social sciences and humanities. Using Qualtrics email distribution, we sent an email inviting them to participate in the study with a link to the survey. Upon clicking the link, they begin by giving their consent and answering a filter question about whether they have ever used GenAI; if they say no, we ask questions about why, current perceptions, concerns about AI, and any future intention, followed by some demographics questions then the survey is terminated. If they said yes, they have used GenAI, they continue to the main survey, where they start by reporting their comfort with technology, their use of GenAI technology in teaching, research, and service, and their views and concerns about GenAI. Then, it covered the main TAM theoretical constructs, followed by some demographic questions. Also, a couple of attention-check questions were added to the survey to ensure respondents were paying sufficient attention and avoid speeding through the survey or giving random answers that could harm data quality. Data collection took place from January to March 2024.

Measures

All the theoretical constructs in this study used existing measures, adapted to fit the AI context and measured using multi-item scales validated in previous research.
Behavior Intention, is the intention level to adopt GenAI, was measured using a five-point likert scale adapted from (Youk & Park, 2023) with four statements such as “I will be willing to use the AI technology in the future” .
Perceived Usefulness. The degree to which a person believes that using GenAI technology would enhance his or her job performance (Davis et al., 2024). This construct was measured using a five-point Likert scale adapted from both (Holden & Rada, 2011; Youk & Park, 2023) with seven statements such as “Using the Gen. Technology improves my performance in my job” ( ).
Perceived ease of use, is the degree to which a person believes that using GenAI would be free from effort, was measured with a nine-item, five-point Likert scale adapted from (Holden & Rada, 2011) including statements such as “I find it easy to get the technology to do what I want it to do” .
AI self-efficacy is the users’ personal confidence towards successfully and purposefully using the GenAI technology. The scale is adopted from (Holden & Rada, 2011). The construct used a ten-point Guttman scale ( not at all confident to totally confident) such as “I could complete any desired task using GenAI technology if there was no one around to tell me what to do as I go.” .
Attitudes toward GenAI, refers to an individual’s overall affective and cognitive evaluation of a new technology. This was measured with a four-item using a fivepoint Likert scale adapted from (Choung et al., 2022) including statements such as “I feel positive toward GenAI” .
Social Reinforcement refers to the extent that technology was encouraged by others within their reference group or social networks. This was measured with a seven-item using a five-point Likert scale adapted from (Compeau & Higgins, 1995) and composed of five statements such as Respondents were asked to assess to what extent the use of GenAI was encouraged by “your job, manager, peers, friend, family, co-workers, subordinates.” ( ).
Trust assesses the psychological state of having positive expectations about GenAI (Choung et al., 2022). We adapted a scale from Huh et al., (2005) that asked respondents to rate GenAI on four items using a five-point scale (believable, trustworthy, credible, and reliable) ( ).
Comfort with technology measures the acceptance, use and overall level of comfort with technology. Here, it was measured using a five-point Likert scale adapted from (Rosen et al., 2013) that consisted of six items such as “I feel that I get more accomplished because of technology” ( ).
Finally, AI concerns included nine items obtained from media reports and discussions with many practitioners. Each item was assessed on a five-item scale and the measure was obtained by averaging the score of the nine items (e.g., plagiarism, copyright, accountability, originality etc.) .

Sample characteristics

A total sample size of 294 higher-education faculty completed the survey. The sample consisted of Full Professors ( ), Associate Professors ( ), Assistant Professors , Faculty in residence ( ), Adjunct Faculty ( ), and Visiting Faculty ( ). Of the sample, did not hold an administrative position. They had multiple years of experience, for years of experience, between 9 and 14 years, between 4 and 8 years, between 15 and 20 years. Participants were Females ( ), Males , Prefer not to answer ( ), and Non-binary/others ( ). Participants identified as Caucasian (69%), Black or African American (7%), Asian or Pacific Islander (7%), Hispanic ( ), Mixed race and others ( ). Their age ranged from . Only of college professors have reported that they used GenAI before.

Research questions

RQ1: Does the TAM model perform similarly with GenAI as it does with other technologies?
RQ2: What is the relationship between the TAM model (PU and PEU) and the SCT (self-efficacy and social reinforcement) in understanding GenAI adoption?
RQ3: How does Trust influence the relationship between the TAM model (PU and PEU) and the SCT (self-efficacy and social reinforcement) in understanding GenAI adoption?

Results

A series of statistical analyses were conducted to test the study hypotheses. The first research question involves evaluating the hypotheses related to the TAM with GenAI technology. For the first hypothesis, a multiple regression was run to predict College Professors’ attitudes toward using GenAI from (a) Perceived Usefulness (PU) and (b) Perceived Ease of Use (PEU). Results found that both variables statistically significantly predicted the attitude to use , accounting for of the variance. The PU and PEU were found to positively predict the attitude to use GenAI. Hence, H1 (a,b) was fully supported (see Table 1).
Similarly, for the second hypothesis, a multiple regression was run to predict College Professors’ behavioral intention to use GenAI from (a) PU, (b) PEU, and (c) attitudes. Results indicated that the model explained of the variance and that the model was a significant predictor of behavioral intention to use , . Only the PU and attitude , were found to positively predict behavioral intention to use
Table 1 Multiple regression with TAM
Attitude Behavioral Intention
Sig Sig
Perceived usefulness 0.655 0.000 0.347 0.000
Perceived ease of use 0.163 0.003 0.645
Attitude 0.462 0.000
GenAI, but PEU was not significant ( ). Hence, H2 (a, c) were only supported (see Table 1).
The third hypothesis predicted College Professors’ PU of using GenAI from PEU. Results found that the model was significant , accounting for of the variance, and PEU was found to positively predict Perceived Usefulness to use GenAI. Hence, H3 was supported.
To test the H4 hypothesis, we conducted a mediation analysis using the SPSS PROCESS macro (Hayes, 2017) to examine whether attitudes mediate the relationship between (a) PU and (b) PEU on college Professors’ behavior intention to use GenAI. Using model 4, we ran a bootstrap analysis with the default 5000 iterations. For PU, results indicated that the model was a significant predictor of attitude [ F(2, 188) with accounting for the variance. Results showed that attitudes mediated the relation between PU and behavior intention ( CI of 0.24 to 0.56 ). While for PEU, results indicated that the model was a significant predictor of attitude with accounting for the variance. Results showed that attitudes mediated the relation between PEU and behavior intention ( of 0.27 to 0.58 ). Thus, only H4 (a,b) was fully supported (see Fig. 2).
The second research question and its corresponding hypotheses aims to integrate and examine the TAM model with SCT theory. The fifth hypothesis is about self-efficacy. A multivariate multiple regression was run to assess how self-efficacy influences Professors’ (a) PEU, (b) attitudes and (c) behavioral intention to use GenAI. Results found that GenAI Self-efficacy statistically significantly predicted PEU ( CI [0.090, 0.172]), but neither predicted attitude ( ), nor behavioral intention to use GenAI ( ). Thus, only H5a is supported (see Table 2).
For the sixth hypothesis, a multivariate multiple regression was run to examine the influence of social reinforcement on PU, PEU, attitude and self-efficacy. Results found that Social Reinforcement statistically significantly predicted all the variables; PU CI , PEU CI , ), attitude ( ), and self-efficacy ( , CI [0.013, 0.702]). Thus, all H6 (a,b,c,d) was fully supported (see Table 2).
To test the seventh hypothesis, we conducted a mediation analysis using the SPSS PROCESS macro (Hayes, 2017) to examine whether social reinforcement mediates the relationship between (a) PEU and (b) PU and College Professors’ behavioral intention to use GenAI. Using model 4, we ran a bootstrap analysis with the default 5000 iterations. For PEU, results showed that the model was significant , and that social reinforcement mediated the relation between PEU and
Fig. 2 TAM Mediation Analysis (H4)
Table 2 Comparison of Factors affecting Technology Adoption
Self-efficacy Social reinforcement Trust
B Sig Sig Sig
Perceived usefulness 0.551 0.348 0.000 0.525 0.000
Perceived ease of use 0.131 0.000 0.161 0.003 0.316 0.000
Attitude 0.113 0.335 0.000 0.630 0.000
Behavior intention 0.571 0.238 0.003 0.503 0.000
Trust 0.072 0.011 0.232 0.001
Self-efficacy 0.357 0.042 0.466 0.011
Social reinforcement 0.061 0.042 0.254 0.001
Fig. 3 Social Reinforcement Mediation Analysis (H7)
behavior intention ( of 0.00 to 0.12 ) (See Fig. 3). However, for PU, mediation was not established because the regression coefficient between social reinforcement and behavior intention was not significant at ( ) and the Zero falls within the confidence interval range ( -0.07 to 0.03 ). This is explained by the direct effect ( ) being nearly identical to the total effect ( ), suggesting there was no indirect effect. Hence, H7a was only supported.
The third research question and its corresponding hypotheses address trust perception. For the eighth hypothesis, a multivariate multiple regression found that Trust statistically significantly influenced PU ( ), PEU ( CI ), attitude ( CI [0.493, ), and behavioral intention to use GenAI ( CI [ 0.354 , 0.652]), as shown in Table 2. Thus, H8 (a,b,c,d) was fully supported (see Table 2).
Since Trust was one of the strong factors that influenced the TAM model, and to answer the ninth hypothesis, we conducted a moderation analysis using model 1 with the SPSS PROCESS macro (Hayes, 2017) to assess whether Trust moderates the relation between PU and behavioral intention towards GenAI. For PU, results indicated that the overall model was significant . The interaction between PU and Trust was significant ( ), indicating that the relationship between PU and behavior intention depends on the level of Trust. Analysis revealed that the simple slopes at different levels of Trust (e.g., low, medium, and high levels) are all positive, indicating that the PU has a positive effect on the behavior intention at all levels of Trust. However, the magnitude of this positive effect decreases as the Trust increases. Thus, the relationship between the PU on behavior intention was stronger when the Trust was low ,
Fig. 4 Moderation with Trust
Fig. 5 Mediation with Trust
compared to when it was high . Thus, H9 was supported (see Fig. 4).
For hypothesis ten, we conducted a mediation analysis using the SPSS PROCESS macro (Hayes, 2017) to examine whether Trust mediates the relationship between social reinforcement and behavior intention to use GenAI. Using model 4, results indicated that the model was significant and that trust mediated the relation between social reinforcement and behavior intention ( of 0.04 to 0.19 ). It is worth noting that only the indirect effect was significant, while the direct effect was not significant ( ). Hence, H10 was supported (see Fig. 5).

Discussion

By integrating the TAM model and SCT theory, this study explored higher education faculty’s perceptions, attitudes, and experiences regarding GenAI. It also examines the potential benefits and drawbacks, motivations and concerns, and factors behind adopting or rejecting GenAI tools. Overall, our results offer theoretical support for the proposed model to evaluate GenAI adoption decisions among college processors (see Table 3).
Table 3 Study hypotheses summary
Hypothesis number Hypothesis paths/ relations Beta P value Result
H1-TAM PU → Att 0.655 Supported
PEU → Att 0.163 Supported
H2-TAM 0.347 Supported
PEU → BI Not supported
Att → BI 0.462 Supported
H3-TAM PEU → PU 0.367 Supported
H4-Mediation Att mediates
0.39 Supported
PEU → BI 0.41 Supported
H5-Self-Efficacy (SE) SE → PEU 0.131 Supported
SE → Att Not Supported
SE → BI Not Supported
H6-Social Reinforcement (SR) 0.348 Supported
SR → PEU 0.161 Supported
SR → Att 0.335 Supported
SR → SE 0.357 Supported
H7- Mediation SR mediates
PEU → BI 0.05 Supported
Not Supported
H8-Trust (T) 0.525 Supported
T → PEU 0.316 Supported
Att 0.630 Supported
0.503 Supported
H9-Moderation Ton
-0.21 Supported
H10-Mediation T mediate
0.11 Supported

TAM

All hypotheses derived from the Technology Acceptance Model (TAM) were found to be statistically significant. Results indicated that perceived usefulness exerts a considerably stronger influence on attitudes and, specifically, on behavioral intentions compared to perceived ease of use. This highlights the importance of demonstrating the practical benefits of GenAI technology to encourage its adoption.
Notably, while PEU did not demonstrate a direct significant effect on college professors’ behavioral intentions to use GenAI ( ), it exhibited a significant indirect mediation effect on behavioral intentions through attitudes. Specifically, mediation analysis revealed that of the total effect ( ), the indirect effect accounted for , whereas the direct effect was relatively minor, at ( ). Thus, the indirect effect predominantly explains the relationships within the model. These findings suggest that perceptions of ease of use of GenAI alone are insufficient to drive adoption among professors, but when these perceptions are coupled with positive attitudes toward GenAI, they can significantly enhance the likelihood of adoption. This indicates that professors may need compelling reasons and motivations to engage
with GenAI. In contrast, perceived usefulness (PU) demonstrates a strong influence on behavioral intentions to use GenAI, both directly and indirectly through attitudes.

Trust

Overall, trust emerges as the most significant predictor influencing all factors (PU, PEU, SR, SE, Att) related to the decision to adopt GenAI technology, closely followed by social reinforcement. In this study, trust refers to perceptions of believability, credibility, reliability, and trustworthiness. Findings show that trust in technology has a direct effect on Professors’ belief that it will enhance their performance, i.e., perceptions of usefulness. It can also foster a sense of simplicity, which, in turn, boosts their self-efficacy-their confidence in their ability to effectively utilize the technology. Increased trust can lead to more positive attitudes and stronger intentions to adopt GenAI technology. This is particularly relevant in light of the prevailing concerns surrounding this controversial technology (Gupta et al., 2024); professors’ level of trust can dispel their concerns and significantly shape their perceptions, attitudes, and subsequent decisions regarding its adoption. Thus, educational institutions need to offer clear information while addressing any misinformation to alleviate fear and skepticism about GenAI and build trust.
Moderation analysis revealed that trust significantly influences the relationship between perceived usefulness and the intention to adopt GenAI technology. However, this effect is conditional, with the impact becoming weaker at higher levels of trust. This suggests that when professors have a high level of trust in the technology, they may rely less on its perceived usefulness or prioritize their trust over perceived usefulness to determine their intention to adopt it. Conversely, when trust is low, professors will need clearer and more compelling reasons for adoption, leading them to rely heavily on perceptions of usefulness as a motivating factor to encourage them to adopt the technology. Overall, this highlights trust as a strong factor that can potentially overshadow other essential factors like perceived usefulness.

Social reinforcement

The results indicate that social reinforcement serves as another significant predictor for most factors that affect the decisions to adopt, particularly because social reinforcement is a common factor across the Social Cognitive Theory (Bandura, 1999) and various technology adoption theories (Granić & Marangunić, 2019; Venkatesh et al., 2016). This finding is consistent with similar investigations in other fields such as healthcare (Roppelt et al., 2024) and journalism (Trang et al., 2024). This underscores the critical role of social interactions and social environments in shaping professors’ perceptions, attitudes, and decisions regarding the adoption of GenAI technology. Specifically, results found that social reinforcement has a direct and positive influence on perceptions of usefulness and ease of use, fostering favorable attitudes and encouraging the overall decision to adopt the GenAI technology. Additionally, reinforcement from one’s social circles can positively influence professors’ belief in their own ability to use and adopt the GenAI (self-efficacy); the belief that “if others can do it, so can I” serves to motivate and facilitate adoption among peers. Therefore, these findings suggest that social dynamics play a big role in decisions to adopt GenAI technology. This effect is important given that GenAI technology is associated with various ethical concerns, prompting many individuals to
approach it with caution. Thus, social interactions with peers, conversations and positive feedback and experiences shared by others within their social groups, and discussion within the departmental, college and university levels can all influence perceptions, attitudes and use of the technology. This suggests that expectations and behaviors of a social group can influence one’s disposition toward GenAI technology and its perceived value. If using a particular technology becomes common within a community, individuals are more likely to adopt it.
Additionally, mediation analysis showed that with social reinforcement as a mediator, there was no mediating indirect effect between perceived usefulness and behavior intention, only a direct effect. This means that once professors perceive GenAI technology to be useful, they will use it. They do not need further reinforcement from their social networks. This explains why the direct effect was nearly identical to the total effect. On the other hand, for perceived ease of use, there was a significant mediation effect on behavioral intention to use GenAI. This suggests that perceptions of easiness are amplified when reinforced by others within social circles who have used it or hold positive attitudes towards the new technology, leading to a higher likelihood of adoption.

Trust and social reinforcement

The mediation analysis showed there was no direct effect from social reinforcement in behavioral intention, but there was a significant indirect effect of trust mediating the relationship between social reinforcement and behavioral intention. This means that while social reinforcement alone does not directly influence intention, it enhances trust, which subsequently increases the likelihood of technology adoption. This is because such social interactions can be a validation and reinforcement that enhance one’s trust, as well as it can help reduce uncertainty or alleviate doubts, thus feeling more confident to use it. However, simply receiving social reinforcement does not directly lead to an intention to use the technology because it depends on the credibility of the source and pre-existing views that can make them skeptical and more resistant to positive views shared by others.

Self-efficacy

The results indicated that self-efficacy can only influence perceptions of ease of use, which is in line with past research showing a positive relationship between technology self-efficacy and perceived ease of use (Holden & Rada, 2011). This is logical because more confidence in one’s ability to use GenAI will lead to perceived mastery and ease of use. However, results also indicated that this confidence alone is insufficient to generate positive attitudes or behavioral intentions to use GenAI, suggesting that the issue may not lie in the ability to use the technology but rather in the ongoing concerns and controversies surrounding its use. This finding aligns with past research that the relationship between self-efficacy and behavioral intentions is often mixed and not always significant (Holden & Rada, 2011; Motshegwe & Batane, 2015). Attitudes and behavioral intentions are shaped by a broader set of cognitive, social, and contextual factors such as values, beliefs, institutional support, trust, and peer influence, which may overshadow the influence of self-efficacy alone. According to Bandura (1997), four factors impact self-efficacy in different ways: mastery of experiences, social modeling, verbal persuasion, and
emotional responses. These factors highlight the complexity of self-efficacy, with social reinforcement and trust playing significant roles in shaping it. This suggests that self-efficacy, while important, is not the sole driver of attitudes or behavioral intentions, which are instead shaped by a much wider range of considerations.
Another possible explanation for the limited significance of self-efficacy, despite faculty feeling capable of using AI technology (i.e., high self-efficacy), they may question its relevance or effectiveness. This skepticism can undermine their intentions to use the technology, which helps explain why perceived usefulness had a stronger impact on attitudes and behavioral intentions than ease of use. Additionally, resistance may also arise from faculty being accustomed to traditional teaching methods, concerns about increased workload or time commitments to use the AI technology, or perceived threats to autonomy. Therefore, while faculty may have the self-efficacy to use AI, they may still hold negative attitudes or fail to integrate it into their work.

Conclusion

This study integrated the TAM model and SCT theory to explore higher education faculty’s perceptions, attitudes, motivations, and concerns and examines the potential benefits, drawbacks, and factors behind adopting or rejecting GenAI tools. Results found theoretical support for the proposed model to evaluate GenAI adoption. Specifically, findings show that communicating the practical benefits of GenAI is necessary for GenAI adoption. However, perceived ease alone is insufficient to encourage adoption, but when coupled with positive attitudes and social circles, it becomes influential in encouraging adoption. Furthermore, trust and social reinforcement are major and critical players in influencing GenAI adoption decisions. There is a social dynamic involved in the GenAI adoption process that can shape professors’ perceptions, attitudes, and decisions. Positive social interactions can enhance trust by acting as a validation or reinforcement of one’s views, as well as can reduce uncertainty or alleviate doubts, which increases the likelihood of technology adoption.

Implications, future studies, and limitations

This study explores the perspective of higher education faculty regarding the utilization and adoption of GenAI. A deeper understanding of educators’ perspectives can shape higher education policies and guide university leaders, policymakers and educators in enhancing the educational experience for college students in an AI-driven workforce. As trust plays a pivotal role in the adoption of GenAI, educational institutions need to provide transparent information and address misinformation to reduce fear and skepticism surrounding the technology, while effectively communicating its practical benefits and functionalities to build credibility. Given the significant role of social dynamics in shaping faculty attitudes toward GenAI, constructive social dialogue and shared experiences within departments, colleges, and universities can greatly influence perceptions of new technologies. Thus, fostering a supportive social environment can deepen our understanding of the potential impacts of GenAI in educational settings, contributing to a more informed and equitable approach to improving educational practices and outcomes.
Future studies could further explore the role of institutional support and messaging regarding the use of AI. Although self-efficacy had a minimal impact on faculty adoption decisions, further investigation into the underlying reasons is necessary, as it can influence the effective utilization of this technology and provide a better understanding of faculty behaviors. Given the substantive ethical issues surrounding GenAI, it is a technology that will likely follow a less typical adoption path, particularly in higher education. For example, uncertainty regarding the faculty’s use of AI for writing that does not ‘cross the line’ will cause many to pause. Therefore, it is crucial to address concerns related to the responsible and ethical use of GenAI. Institutions also face ongoing decisions regarding the availability of AI tools that are often embedded in existing software. Additionally, using qualitative methods offers a deeper understanding of how faculty members perceive GenAI technology, its ethical considerations, and its potential impact on their teaching practices. Incorporating open-ended questions within the different areas of use could further elucidate the findings. Research should examine how the adoption of GenAI varies across academic disciplines and cultural contexts, identifying unique challenges and opportunities within each field. Given the rapid developments in GenAI, longitudinal research is necessary to assess how perceptions and uses of GenAI evolve over time.
The targeted sampling strategy of non-STEM fields partially limits the generalizability of this study. There is little doubt that educators beyond the scope of this population will differ in their views of the role of GenAI in their work. Furthermore, the study was conducted with two mid-sized research-intensive public universities in the U.S that may have limited generalizability to other educational contexts compared to smaller institutions, private universities, teaching-intensive and international settings, potentially leading to different outcomes. The reliance on self-report data regarding potentially controversial topics may result in responses that might be viewed as more favorable at the expense of accuracy. Additionally, the study’s focus on a specific point in time may not capture evolving attitudes or changes in behaviors, especially as GenAI technology rapidly advances. Another potential limitation is self-selection bias can influence the findings and limit the generalizability of the results. Consequently, the results may not fully reflect the diversity of perspectives, behaviors, or experiences within the broader population. Finally, while we studied the role of trust and social reinforcement, several factors can affect faculty AI adoption such as institutional support and culture, training and professional development, and discipline or areas of specialization.

Appendix

See Tables 4, 5, 6, 7, 8, 9 and 10
Table 4 List of abbreviations
Abbreviation Term
GenAI Generative artificial intelligence
PU Perceived usefulness
PEU Perceived ease of use
TAM Technology acceptance model
UTAUT Unified theory of acceptance and use of technology
SCT Social cognitive theory
BI Behavioral intention
Att Attitude
SR Social Reinforcement
SE Self-efficacy
Table 5 TAM and attitude
Model Unstandardized coefficients Standardized coefficients t Sig
B Std. error Beta
(Constant) 0.226 0.268 0.844 0.400
PU 0.749 0.061 0.655 12.221 0.000
PEU 0.235 0.077 0.163 3.040 0.003
a. Dependent variable: attitude
Table 6 TAM and behavioral intentions
Model Unstandardized coefficients Standardized coefficients Beta t Sig
B Std. error
(Constant) 1.139 0.258 4.424 0.000
PEU 0.035 0.076 0.024 0.462 0.645
PU 0.399 0.079 0.347 5.065 0.000
Attitude 0.464 0.070 0.462 6.633 0.000
a. Dependent variable: Behavioral_Intention
Table 7 PU and PEU
Model Unstandardized coefficients Standardized coefficients t Sig
B Std. error Beta
1 (Constant) 1.613 0.298 5.414 0.000
PEU 0.471 0.086 0.367 5.487 0.000
Table 8 Self-efficacy
Dependent variable Parameter B Std. Error t Sig 95% Confidence interval
Lower bound Upper bound
PU Intercept 3.116 0.199 15.639 0.000 2.723 3.509
Self_Efficacy 0.017 0.029 0.597 0.551 -0.040 0.074
PEU Intercept 2.538 0.143 17.710 0.000 2.255 2.820
Self_Efficacy 0.131 0.021 6.334 0.000 0.090 0.172
Att Intercept 3.102 0.226 13.700 0.000 2.655 3.548
Self_Efficacy 0.052 0.033 1.593 0.113 -0.012 0.117
BI Intercept 4.020 0.229 17.569 0.000 3.569 4.471
Self_Efficacy 0.019 0.033 0.568 0.571 -0.046 0.084
SR Intercept 2.221 0.205 10.816 0.000 1.816 2.626
Self_Efficacy 0.061 0.030 2.046 0.042 0.002 0.119
Trust Intercept 2.330 0.195 11.944 0.000 1.945 2.715
Self_Efficacy 0.072 0.028 2.564 0.011 0.017 0.128
a. Computed using alpha
Table 9 Social reinforcement
Dependent variable Parameter B Std. error t Sig 95% Confidence interval
Lower bound Upper bound
PU Intercept 2.317 0.181 12.797 0.000 1.960 2.674
Social_Reinforcement 0.348 0.065 5.345 0.000 0.220 0.477
PEU Intercept 2.973 0.150 19.797 0.000 2.677 3.269
Social_Reinforcement 0.161 0.054 2.984 0.003 0.055 0.268
Att Intercept 2.566 0.211 12.136 0.000 2.149 2.983
Social_Reinforcement 0.335 0.076 4.400 0.000 0.185 0.485
BI Intercept 3.520 0.218 16.158 0.000 3.090 3.949
Social_Reinforcement 0.238 0.078 3.037 0.003 0.083 0.393
Trust Intercept 2.194 0.187 11.701 0.000 1.824 2.564
Social_Reinforcement 0.232 0.067 3.447 0.001 0.099 0.366
SE Intercept 5.610 0.486 11.553 0.000 4.652 6.568
Social_Reinforcement 0.357 0.175 2.046 0.042 0.013 0.702
Table 10 Trust
Dependent variable Parameter B Std. error t Sig 95% Confidence interval
Lower bound Upper bound
PU Intercept 1.756 0.183 9.580 0.000 1.395 2.118
Trust 0.525 0.062 8.428 0.000 0.402 0.648
PEU Intercept 2.509 0.156 16.074 0.000 2.201 2.817
Trust 0.316 0.053 5.961 0.000 0.212 0.421
Att Intercept 1.676 0.205 8.158 0.000 1.271 2.081
Trust 0.630 0.070 9.024 0.000 0.493 0.768
BI Intercept 2.732 0.222 12.294 0.000 2.294 3.170
Trust 0.503 0.076 6.664 0.000 0.354 0.652
SR Intercept 1.905 0.217 8.773 0.000 1.477 2.333
Trust 0.254 0.074 3.447 0.001 0.109 0.400
SE Intercept 5.241 0.534 9.810 0.000 4.187 6.295
Trust 0.466 0.182 2.564 0.011 0.107 0.824
a. Computed using alpha

Acknowledgements

Not applicable.

Author contributions

AS: conception, design of the work, acquisition, analysis, interpretation of data, drafted the work, Approved the submitted version (and any substantially modified version that involves the author’s contribution to the study); Agreed both to be personally accountable for the author’s own contributions and to ensure that questions related to the accuracy or integrity of any part of the work, even ones in which the author was not personally involved, are appropriately investigated, resolved, and the resolution documented in the literature. KH: design of the work, acquisition, interpretation of data, drafted the work, Approved the submitted version (and any substantially modified version that involves the author’s contribution to the study); Agreed both to be personally accountable for the author’s own contributions and to ensure that questions related to the accuracy or integrity of any part of the work, even ones in which the author was not personally involved, are appropriately investigated, resolved, and the resolution documented in the literature.

Funding

The research was not supported by external funding organizations.

Availability of data and material

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

This manuscript has not been published or submitted for review to any other outlet. The study was reviewed and approved by the University of Nevada, Las Vegas Office of Research Integrity.

Competing interests

The authors declare that they have no financial or non-financial competing interests.
Received: 25 October 2024 Accepted: 10 February 2025
Published online: 14 March 2025

References

Bandura, A. (1997). Self-efficacy: The exercise of control. Freeman.
Bandura, A. (1999). Social cognitive theory: An agentic perspective. Asian Journal of Social Psychology, 2(1), 21-41. https:// doi.org/10.1111/1467-839X. 00024
Bandura, A. (2002). Social cognitive theory in cultural context. Applied Psychology, 51(2), 269-290. https://doi.org/10. 1111/1464-0597.00092
Barrett, A., & Pack, A. (2023). Not quite eye to A.I.: Student and teacher perspectives on the use of generative artificial intelligence in the writing process. International Journal of Educational Technology in Higher Education, 20(1), 59. https://doi.org/10.1186/s41239-023-00427-0
Bower, M., Torrington, J., Lai, J. W. M., Petocz, P., & Alfano, M. (2024). How should we change teaching and assessment in response to increasingly powerful generative Artificial Intelligence? Outcomes of the ChatGPT teacher survey. Education and Information Technologies. https://doi.org/10.1007/s10639-023-12405-0
Buskens, V. (2020). Spreading information and developing trust in social networks to accelerate diffusion of innovations. Trends in Food Science Technology, 106, 485-488. https://doi.org/10.1016/j.tifs.2020.10.040
Chiu, T. K. F. (2024). Future research recommendations for transforming higher education with generative AI. Computers and Education: Artificial Intelligence, 6, 100197. https://doi.org/10.1016/j.caeai.2023.100197
Choung, H., David, P., & Ross, A. (2022). Trust in AI and its role in the acceptance of AI technologies. International Journal of Human-Computer Interaction, 39(9), 1727-1739. https://doi.org/10.1080/10447318.2022.2050543
Compeau, D. R., & Higgins, C. A. (1995). Application of social cognitive theory to training for computer skills. Information Systems Research, 6(2), 118-143. https://doi.org/10.1287/isre.6.2.118
Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. M/S Quarterly, 13(3), 319-340. https://doi.org/10.2307/249008
Davis, F. D., Bagozzi, R. P., & Warshaw, P. R. (1989). User acceptance of computer technology: A comparison of two theoretical models. Management Science, 35(8), 982-1003. https://doi.org/10.1287/mnsc.35.8.982
Davis, F. D., Granić, A., & Marangunić, N. (2024). The technology acceptance model: 30 years of TAM. Springer International Publishing AG.
Granić, A., & Marangunić, N. (2019). Technology acceptance model in educational context: A systematic literature review. British Journal of Educational Technology, 50(5), 2572-2593. https://doi.org/10.1111/bjet. 12864
Gupta, R., Nair, K., Mishra, M., Ibrahim, B., & Bhardwaj, S. (2024). Adoption and impacts of generative artificial intelligence: Theoretical underpinnings and research agenda. International Journal of Information Management Data Insights, 4(1), 100232. https://doi.org/10.1016/j.jjimei.2024.100232
Hartley, K., Hayak, M., & Ko, U. H. (2024). Artificial intelligence supporting independent student learning: An evaluative case study of ChatGPT and learning to code. Education Sciences, 14(2), 120.
Hayes, A. F. (2017). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach. Guilford publications.
Holden, H., & Rada, R. (2011). Understanding the influence of perceived usability and technology self-efficacy on teachers’ technology acceptance. Journal of Research on Technology in Education, 43(4), 343-367. https://doi.org/ 10.1080/15391523.2011.10782576
Huh, J., DeLorme, D. E., & Reid, L. N. (2005). Factors affecting trust in on-line prescription drug information and impact of trust on behavior following exposure to DTC advertising. Journal of Health Communication, 10(8), 711-731.
Labadze, L., Grigolia, M., & Machaidze, L. (2023). Role of AI chatbots in education: Systematic literature review. International Journal of Educational Technology in Higher Education, 20(1), 56. https://doi.org/10.1186/ s41239-023-00426-1
Lim, W. M., Gunasekara, A., Pallant, J. L., Pallant, J. I., & Pechenkina, E. (2023). Generative AI and the future of education: Ragnarök or reformation? A paradoxical perspective from management educators. The International Journal of Management Education, 21(2), 100790. https://doi.org/10.1016/j.jime.2023.100790
Marikyan, D., Papagiannidis, S., & Stewart, G. (2023). Technology acceptance research: Meta-analysis. Journal of Information Science. https://doi.org/10.1177/01655515231191177
McClain, C. (2024). Americans’ use of ChatGPT is ticking up, but few trust its election information. Pew Research Center. https://www.pewresearch.org/short-reads/2024/03/26/americans-use-of-chatgpt-is-ticking-up-but-few-trust-its-election-information/
Mollick, E. (2024). Co-Intelligence: Living and Working with AI. Penguin Publishing Group. http://ebookcentral.proquest. com/lib/unlv/detail.action?docID=30848006
Motshegwe, M., & Batane, T. (2015). Factors influencing instructors’ attitudes toward technology integration. Journal of Educational Technology Development and Exchange (JETDE). https://doi.org/10.18785/jetde.0801.01
Pan, X. (2020). Technology acceptance, technological self-efficacy, and attitude toward technology-based self-directed learning: Learning motivation as a mediator. Frontiers in Psychology. https://doi.org/10.3389/fpsyg.2020.564294
Roppelt, J. S., Kanbach, D. K., & Kraus, S. (2024). Artificial intelligence in healthcare institutions: A systematic literature review on influencing factors. Technology in Society, 76, 102443. https://doi.org/10.1016/j.techsoc.2023.102443
Rosen, L. D., Whaling, K., Carrier, L. M., Cheever, N. A., & Rokkum, J. (2013). The media and technology usage and attitudes scale: An empirical investigation. Computers in Human Behavior, 29(6), 2501-2511.
Shieh, A., Tran, B., He, G., Kumar, M., Freed, J. A., & Majety, P. (2024). Assessing ChatGPT 4.0’s test performance and clinical diagnostic accuracy on USMLE STEP 2 CK and clinical case reports. Scientific Reports, 14(1), 9330. https://doi.org/10. 1038/s41598-024-58760-x
Steiss, J., Tate, T., Graham, S., Cruz, J., Hebert, M., Wang, J., Moon, Y., Tseng, W., Warschauer, M., & Olson, C. (2024). Comparing the quality of human and ChatGPT feedback of students’ writing. Learning and Instruction, 91, 101894. https:// doi.org/10.1016/j.learninstruc.2024.101894
Tanantong, T., & Wongras, P. (2024). A UTAUT-based framework for analyzing users’ intention to adopt artificial intelligence in human resource recruitment: A case study of Thailand. Systems, 12(1), 28.
Trang, T. T. N., Chien Thang, P., Hai, L. D., Phuong, V. T., & Quy, T. Q. (2024). Understanding the adoption of artificial intelligence in journalism: An empirical study in vietnam. SAGE Open, 14(2), 21582440241255240. https://doi.org/10.1177/ 21582440241255241
Venkatesh, V., Thong, J. Y. L., & Xu, X. (2016). Unified Theory of Acceptance and Use of Technology: A Synthesis and the Road Ahead (SSRN Scholarly Paper 2800121). https://papers.ssrn.com/abstract=2800121
Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User acceptance of information technology: Toward a unified view. MIS Quarterly, 27(3), 425-478. https://doi.org/10.2307/30036540
Wu, K., Zhao, Y., Zhu, Q., Tan, X., & Zheng, H. (2011). A meta-analysis of the impact of trust on technology acceptance model: Investigation of moderating influence of subject and context type. International Journal of Information Management, 31(6), 572-581. https://doi.org/10.1016/j.jijnfomgt.2011.03.004
Youk, S., & Park, H. S. (2023). Who is (communicatively more) responsible behind the wheel? Applying the theory of communicative responsibility to TAM in the context of using navigation technology. Human-Machine Communication, 6(1), 11.
Yusuf, A., Pervin, N., & Román-González, M. (2024). Generative AI and the future of higher education: A threat to academic integrity or reformation? Evidence from multicultural perspectives. International Journal of Educational Technology in Higher Education, 21(1), 21. https://doi.org/10.1186/s41239-024-00453-6

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

  1. © The Author(s) 2025. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creativecommons.org/licenses/by/4.0/.
  2. a. Dependent variable: PU