Canadian Academy Discover, Publish, Thrive
The Chat Generative Pre-trained transformer, known as ChatGPT, holds promise for transforming higher education, but factors influencing its adoption among economics students in Ghanaian institutions are unclear. Utilizing the Unified Theory of Acceptance and Use of Technology (UTAUT2), a study surveyed 306 Ghanaian economics students, revealing design and interactivity significantly impact perceived trust, while factors like social influence and hedonic motivation drive behavioral intentions. These intentions, along with facilitating conditions, significantly influence actual ChatGPT usage. However, ethics and effort expectancy show no significant impact, suggesting implications for integrating AI tools to enhance educational accessibility and inclusivity.
Artificial intelligence (AI) technologies are increasingly seen as vital for enhancing global education and achieving inclusive learning goals. The emergence of AI-based tools, such as the Chat Generative Pre-trained Transformer (ChatGPt), is reshaping education in the 21st century, offering solutions for answering commonly asked questions and providing quick, individualized support to students. However, concerns about maintaining ethical standards in research and avoiding academic dishonesty persist. Despite these concerns, many scholars recognize the advantages of ChatGPt in higher education settings due to its potential to improve learning outcomes and bridge digital gaps. Understanding the factors influencing students’ intentions and usage of ChatGPt, especially in the context of Ghanaian higher education institutions (HEIs), remains a critical area of study. To address this gap, this study employs the Unified Theory of Acceptance and Use of Technology (UTAUT2) framework, supplemented with factors such as design, interactivity, and ethics, to explore students’ perceptions and behaviors towards ChatGPt. Structural Equation Modeling (SEM) and Artificial Neural Network (ANN) analysis are utilized to provide a comprehensive understanding of ChatGPt adoption, offering insights for policymakers to promote quality and inclusive education through AI integration. This paper outlines the theoretical framework, research methods, results, and implications for future studies in this field.
The use of artificial intelligence (AI) is increasingly prevalent across various fields, notably in higher education, where AI-based tools have significantly impacted teaching and learning processes. These tools automate tasks, analyze vast amounts of data, and offer predictive insights, potentially revolutionizing numerous aspects of our lives. ChatGPt, a recent AI-based tool, has emerged as a potential game-changer in higher education, despite concerns about integrity and academic misconduct. It supports instructors in course material preparation, provides suggestions, performs translations, and evaluates student performance. For students, ChatGPt functions as a virtual tutor, aiding in answering questions, summarizing information, and facilitating collaboration. In fields like economics, studies show ChatGPt’s substantial advantages, surpassing average students in standardized tests and assisting in various tasks such as answering queries and improving writing. However, its adoption among economics students in Ghana remains limited, highlighting the need for further research on factors influencing its use. Investigating students’ behavioral intentions and utilization of ChatGPt in higher education provides valuable insights into AI adoption, informing policy directions to enhance learning outcomes, especially in complex subjects like economics. This research can deepen our understanding of AI adoption in Ghanaian higher education and guide the development of strategies to improve academic achievement in fields requiring extensive reading, research, and quantitative data analysis.
This study employs the Extended Unified Theory of Acceptance and Use of Technology (UTAUT2) model to explore the factors influencing the adoption of technologies like ChatGPT. UTAUT2 encompasses seven key constructs to gauge individuals’ behavioral intention towards adopting new technologies. Recent research has highlighted the model’s applicability in various contexts, particularly within information systems. Concerns regarding academic ethics have led to hesitancy in adopting ChatGPT among institutions and students. To address these concerns, the study extends the UTAUT2 model by incorporating additional factors such as design, interactivity, ethics, and perceived trust, especially among economics students in Ghanaian higher education institutions. Rahim et al. (2022) emphasize the significance of design, interactivity, and ethics in shaping individuals’ trust perceptions towards AI-based technologies, hence warranting their inclusion in the extended framework.
Design significantly influences technology’s distinctive qualities, impacting learning efficiency and cognitive load. Subcategories like functions and security can enhance consumers’ trust in technology platforms. Research suggests that students’ trust in ChatGPt is influenced by design, highlighting the need for further exploration, as recent studies in Ghana have overlooked this aspect.
Interactivity, defined as user-platform interaction, is crucial for trust and technology adoption success. Perceived interactivity significantly influences trust in various sectors like e-learning and mobile banking, akin to its impact on ChatGPT’s user engagement. Understanding the link between perceived trust and interactivity is key for comprehending technology adoption behaviors. Studies suggest that perceived interactivity positively affects trust, impacting students’ intention to use ChatGPT. Thus, this study hypothesizes that interactivity significantly influences perceived trust in ChatGPT use by students.
Ethics delineate acceptable from inappropriate behavior, encompassing AI like ChatGPT, which poses ethical dilemmas due to potential biases and privacy concerns. Navigating this ethical domain is crucial for fostering trust in technology, especially in AI applications, as users’ perceptions of ethics influence their trust. Hence, the study hypothesizes that ethics significantly positively impact students’ trust in using ChatGPT.
Technology’s success in overcoming user skepticism is contingent upon establishing trustworthiness. Trust, defined as confidence and reliability in a system, greatly influences technology adoption. Perceived trust in AI-based platforms, including ChatGPT, significantly impacts user intention to utilize them. Hence, we propose the hypothesis that perceived trust positively affects students’ intention to use ChatGPT.
Performance expectancy, a key UTAUT2 concept, relates to how individuals perceive technology use and its impact on outcomes. Studies consistently show that believing technology enhances outcomes influences one’s intention to use it. In this study, students’ views on ChatGPT’s accuracy were linked to their performance expectancy, supporting the hypothesis that it positively influences their intention to use ChatGPT.
Effort expectancy, reflecting the perceived ease of using a system, is crucial for understanding users’ adoption of new technologies like ChatGPt, particularly among economics students. Numerous studies have shown that effort expectancy strongly predicts users’ willingness to embrace AI-based tools. Given the ongoing challenges in ChatGPt’s accuracy, reliability, and ethical concerns, investigating students’ perceptions of effort expectancy is increasingly vital, especially in the context of Ghanaian higher education. Recognizing effort expectancy as pivotal for user-friendly perceptions, this study proposes the hypothesis that it significantly influences students’ behavioral intention to use ChatGPt.
Social influence, as defined by various researchers, plays a crucial role in shaping individuals’ intentions and behaviors towards technology adoption, such as ChatGPT. Studies highlight its positive impact on students’ readiness to embrace AI-based tools like chatbots. In this context, social influence refers to the perception of significant others’ influence, including family, friends, and classmates, on one’s decision to use ChatGPT. Consequently, it is hypothesized that social influence significantly enhances students’ behavioral intention to utilize ChatGPT.
The adoption of technology can be driven by hedonic motivation, stemming from the pleasure and delight associated with using it. This enjoyment transforms tasks into enjoyable experiences, leading to increased satisfaction and a strong intention to continue usage. Previous research indicates that hedonic motivation positively impacts students’ intentions to use AI-based technologies like chatbots. Consequently, this study proposes the hypothesis that hedonic motivation significantly influences students’ behavioral intention to use ChatGPT.
Habit, defined as recurring actions performed automatically over time, is crucial in technology adoption. Studies by Varma & Marler (2013) and Rahim et al. (2022) highlight habit’s influence on technology use intentions, such as chatbot adoption. Almahri et al. (2020) and Strzelecki (2023) further support this, showing significant positive effects of habit on students’ intentions and actual usage of technologies like ChatGPt. Formulated hypotheses posit that habit positively impacts both intention and actual use behavior of ChatGPt among students.
Facilitating conditions encompass individuals’ perceptions of technical infrastructure support for contemporary technologies. These conditions, vital for technology functionality, include software integration and network support. Positive use behavior towards technology is influenced by perceived infrastructure readiness. While previous studies examined this impact on AI technologies like chatbots, there is limited evidence regarding ChatGPT. Thus, this study aims to assess the influence of facilitating conditions on students’ actual use of ChatGPT, hypothesizing a significant positive effect.
Behavioural intention, representing readiness to use technology for various tasks, is often measured by attitudes and beliefs. Attitude, reflecting a psychological inclination towards system evaluation, and beliefs about innovation significantly predict intention. These factors are key drivers of user behavior, culminating in actual technology use. Studies show that behavioral intention strongly influences actual usage, making it a critical factor, as hypothesized regarding students’ use of ChatGPT.
This study utilized the extended UTAUT2 model and a hybrid analytical approach, combining structural equation modeling and artificial neural network (SEM-ANN), to understand factors affecting the behavioral intentions and actual usage of ChatGPT among economics students in Ghanaian higher education institutions. The results show a significant positive impact of design on students’ perceived trust, consistent with previous findings. Interactivity also positively influences perceived trust, indicating that students trust ChatGPT more if they find interaction beneficial and efficient. Perceived trust significantly affects students’ behavioral intentions to use ChatGPT, aligning with previous research on technology adoption.
Performance expectancy, social influence, hedonic motivation, and habits positively influence students’ behavioral intentions to use ChatGPT. Moreover, behavioral intentions significantly impact the actual usage of ChatGPT, confirming previous findings. Facilitating conditions also play a significant role in influencing actual usage. However, effort expectancy and habit do not significantly affect behavioral intentions or actual usage, with potential explanations including methodological differences and cultural or technological disparities in Ghana’s context. These findings contribute to understanding the dynamics of AI-based tool adoption in educational settings.
This study utilized the extended theory of acceptance and use of technology (UTAUT2) to explore the factors influencing students’ behavioral intentions and actual usage of ChatGPT in higher education . Data from 306 economics students in a Ghanaian university were analyzed using hybrid structural equation modeling and artificial neural network (SEM-ANN). Findings revealed that design and interactivity significantly enhance perceived trust in ChatGPT. Additionally, perceived trust, social influence, performance expectancy, and habit positively affect students’ behavioral intentions, with hedonic motivation being the primary driver. Behavioral intentions and facilitating conditions strongly predict actual ChatGPT usage, emphasizing their role in technology adoption. The study underscores the importance of bolstering perceived trust, behavioral intentions, and actual usage by focusing on the identified factors. These insights are particularly relevant for Ghanaian economics students due to the increasing integration of AI in higher education, highlighting the significance of a well-designed ChatGPT interface in fostering trust and reliability in AI tools.
Source:
Iddrisu Salifu, Francis Arthur, Valentina Arkorful, Sharon Abam Nortey & Richard Solomon Osei-Yaw (2024) Economics students’ behavioural intention and usage of ChatGPT in higher education: a hybrid structural equation modelling-artificial neural network approach, Cogent Social Sciences, 10:1, DOI: 10.1080/23311886.2023.2300177