Revolutionizing education: Artificial intelligence empowered learning in higher education

Abstract

Global businesses are actively embracing Industry 4.0 and digital transformation. Simultaneously, the education sector is leveraging digital tools to foster personalized learning and equity. Universities transcending borders and students becoming increasingly global have opened new frontiers through the use of artificial intelligence (AI)-based tools in education. Since the role of AI is inevitable in future education, current research aims to identify the level of awareness of faculty members on the applicability and adoption of artificial intelligence. The study also intended to discover how AI enhanced their learning experience and impacted the degree of work engagement of teachers in higher education. A cluster and multi-stage sampling method was employed to select 250 faculty members from QS (Quacquarelli Symonds) ranked institutions operating in hybrid education modes. Utilizing a quantitative research approach and a structural equation model, factors influencing AI adoption in this context were explored. The findings revealed that AI implementation led to the evolution of robust evaluation and assessment methods, resulting in heightened faculty engagement levels. The study identified that perceived risk, performance expectancy, and awareness play significant roles in influencing work engagement and the adoption of AI within the higher education system through mediating variables, specifically attitude, and behavior.

Keywords:

• Artificial intelligence
• work engagement
• Asia
• higher education
• teachers
• digitalization

1. Introduction

Higher education in the 21st century is evolving rapidly, driven by advances in technology, globalization, and changing student demographics. With the widespread availability of online learning platforms, universities are increasingly offering courses and degree programs that can be completed entirely online (Dieguez et al., 2021). This practice allows more students to access higher education and offers greater flexibility in their learning process (Neumann et al., 2021). Since educational sectors are becoming noticeably more diverse, with students enrolling and learning from a wider range of environments, that leads to a greater emphasis on cross-cultural understanding and global citizenship. As the pace of technological change continues to accelerate, universities are playing a more important role in driving innovation and research (Amornkitpinyo et al., 2021). This leads to more partnerships between academia and industry, as well as a greater focus on entrepreneurship and commercialization. In recent recruitment drive employers are willing to opt for graduates with specific skills and competencies, rather than just a broad-based education. Consequently, notable institutions are transforming towards more skills-based learning patterns that offer students practical, career-focused skills (Kocak et al., 2021). The educational industry is identifying multiple ways to meet the requirements of stakeholders to enhance the quality of higher education (Khan et al., 2022). One of the most optimistic solutions to enhance education is through the implementation of artificial intelligence (AI) (Chedrawi & Howayeck, 2019). The future of artificial intelligence in education is highly promising, as technology is gaining drastic transformation and improving the way we learn and teach (Mishra, 2019).

Artificial intelligence is playing a vital role in upgrading the quality of higher education in numerous ways (Choi, 2020). AI-powered learning approaches have been employed to evaluate students’ performance records, determine their strengths and weaknesses, and provide them with customized learning experiences that are tailored to their individual needs. This approach provides students with a tool kit model to gain knowledge more effectively with a productive outcome (Aldosari, 2020). AI-based technology such as Chatbots, Virtual Assistance tools, and Adaptive Learning Systems offer immersive and engaging learning experiences that allow students to discover complex theories and solutions in a more interactive and meaningful manner (Chaudhry et al., 2023; Pradana et al., 2023). In assessment and feedback, AI assists in grading and appraising student assignments, for example detecting similarities through Turnitin, monitoring students’ participation and involvement in library resource utilization, providing faster and more precise feedback to students, and freeing up time for instructors to focus on other aspects of teaching (Essien et al., 2020). Similarly, AI-powered chatbots offer learners immediate and personalized assistance for their academic and organizational needs, such as answering questions about course materials or providing information on course registration and addressing basic quarries. These systems successfully analyze student data to predict which students are at risk of dropping out or struggling academically. This filtration helps instructors and support staff identify and intervene with students at academic risk early, providing them with the help they need to succeed. It has been identified that various AI applications, for example, Bit.ai, Mendeley, Turnitin, elinik.io, and Coursera tools and platforms support higher education research by analyzing large data sets, generating insights and predictions, and identifying patterns that may be difficult for human researchers to detect (Wenge, 2021).

As technology continues to evolve and improve, we can expect to see even more innovative and effective applications of AI in education, leading to more personalized, engaging, and effective learning experiences for students (Li et al., 2021). This promising journey of AI significantly improves the performance and engagement of teachers in higher education. Since teaching staff in higher education have a critical role both in administrative and academic and are completely engaged in regulatory, accreditation, and other important activities, the adoption of AI can help teachers automate administrative tasks, such as grading assignments, tracking attendance, and providing feedback to students, etc (Bisen et al., 2021). AI also helps teachers identify areas where they can improve their teaching skills and provide personalized professional development opportunities. For example, AI-powered coaching tools provide teachers with feedback on their teaching performance and suggest areas for improvement (Minkevics & Kampars, 2021).

In the landscape of contemporary higher education, persistent challenges such as unequal access, limited inclusivity, and the inadequacy of catering to diverse learning styles pose significant challenges (Odhiambo, 2016). The prevailing use of traditional, one-size-fits-all teaching methods falls short of effectively engaging students with varied learning preferences, hindering the development of active participation and critical thinking skills (Kistyanto et al., 2022). Moreover, reliance on traditional assessment methods fails to capture a comprehensive understanding of students’ knowledge, skills, and practical application, with limited tools for assessing and enhancing non-cognitive skills (Rudolph et al., 2023). Traditional teaching and assessment approaches prove insufficient in addressing these pressing issues. Furthermore, the absence of opportunities for international collaboration and cultural exchange in higher education compounds these challenges. To bridge these gaps, the integration of artificial intelligence (AI) emerges as a pivotal solution. AI algorithms have the capacity to analyze individual learning patterns, allowing for the tailoring of coursework to accommodate diverse preferences. The incorporation of predictive analytics effectively identifies students at risk, facilitating timely interventions to support their academic journey. AI-driven educational content delivery systems offer adaptability to students’ pace, learning styles, and knowledge gaps, revolutionizing the approach to content dissemination. Additionally, the automation of administrative tasks through AI provides the means to liberate faculty and staff from routine responsibilities, allowing them to focus on more impactful activities. Hence, the ongoing research actively contributes to the empowerment of the higher education system through the adoption of AI. The findings of this research endeavor will serve to assist institutional policymakers in recognizing how the adoption of new technology is perceived in higher education, enabling them to provide the necessary infrastructure and training to overcome specific challenges.

Since the role of AI is inevitable in future education, current research aims to identify the level of awareness of faculty members on the applicability and adoption of artificial intelligence. The study also intended to discover how AI enhanced their learning experience and impacted the degree of work engagement and productivity of teachers in higher education. The first section of the manuscript highlights the introduction, and the role of AI in higher education has been discussed, and the second section focuses on existing literature. The third section emphasized the methodological aspects of the research. The results and discussions were presented in the following subsection. The research concluded with suitable practical and theoretical implications.

2. Review of literature

Artificial intelligence has been increasingly integrated into various sectors, including higher education. Previous studies explored the use of AI in higher education, including its applications, challenges, and opportunities.

2.1. Factors that influence attitude toward AI in higher education

Attitude is one of the major concerns to be adapted to any technology or system. The literature, as stated by Al Darayseh (2023), identified the significant impact of AI in education toward the attitude of learners. It is identified that conducive technical advancement and allied infrastructure support the implementation of the new system (Pedral Sampaio et al., 2023). The sense of facility condition in an organization controls the behavioral intention of the workforce (Venkatesh et al., 2012). A study conducted by van Twillert et al. (2020) emphasized the role of facilities and infrastructure that influenced faculties’ attitudes toward adopting Web 2.0 technologies in higher education. These research findings stress the effective facility condition in adopting new technologies in higher education. Therefore, the study proposed the following hypothesis.

H1:

Facility condition significantly influences the attitude of the users in adopting AI in higher education.

Employees’ awareness of new systems and technology has a significant impact on their attitude toward effective adoption. Porter & Graham, (2016) identified how an effective awareness/exploration framework supported institutions to adopt a blended learning approach in higher education. In a similar study, Campillo-Ferrer & Miralles-Martínez, (2021) explored systematic awareness and training programs supported the establishment of flipped classrooms in universities. Similarly, a study conducted by Nikou & Maslov, (2023) notified the role of awareness in the effective implementation of e-learning in higher education during a pandemic. Awareness also played a vital role in implementing technology-integrated courses (Wilson, 2023) and e-learning in higher education (Nyathi & Sibanda, 2022). Considering this outcome and the relationship of awareness in adapting new technology, the following hypothesis is developed.

H2:

Faculties awareness significantly influences the attitude toward adopting AI in higher education.

Perceived risk is the psychological influence on how individuals perceive potential uncertainties about the outcomes and ambiguity associated with the use of technology (Li et al., 2019). According to Gupta & Mathur, (2023), perceived risk significantly influences the adoption of virtual communication by education leaders. Several studies for example (Shin et al., 2017; Wu et al., 2022) identified the perception of faculties while adopting AI in higher education. These studies highlighted how AI interferes with education and the possible risks caused due to the adoption of this technology. The study conducted by Lei et al. (2022) identified the perceived risk to faculties, staff, and students in higher education when introducing robots in higher education. The introduction of AI in educational institutions is expected to bring uncertainty about how it will influence the job roles and responsibilities of academic staff members. This role stress could impact the attitude of faculties towards the adoption of AI in HE. Therefore, the following hypothesis is developed.

H3:

Faculties’ perceived risk negatively and significantly influences their attitude on adopting AI in higher education.

Performance expectancy (PE) has been interpreted as users’ perception of adopting a new system and their attainment of gain in productivity. Abd Aziz et al. (2023) applied the UTAUT framework to identify students’ PE in adopting cloud computing higher education and explored how techno complexity and insecurity impact students’ performance expectancy. A similar study by Yip et al. (2021) identified the adoption of mobile library apps in higher education for learning has increased the performance expectancy of students. Further studies like (Diep et al., 2016; Nikolopoulou et al., 2021) underline the positive performance expectancy of teachers and learners while introducing tech-based teaching pedagogy and forms of tools in higher education. These studies confirm the relationship between PE and attitude toward adopting technology. Therefore, the following hypothesis developed.

H4:

Faculties’ performance expectancy has a significant influence on their attitude toward adopting AI in higher education.

Effort expectancy (EE) is often associated with the amount of easiness or perceived ease of use while adopting the new system (Venkatesh et al., 2003). As Zhu & Huang, (2023) identified factors like a user-friendly interface, easy-to-use design, and clear instructions contribute to a positive attitude in mobile learning. Similar research by Sarfraz et al. (2022) notifies that effort expectancy significantly mediates the learner’s attitude in considering technology-based education and learning approach. Further, Nsamba & Chimbo, (2023) also highlighted that EE has a positive mediating role in adopting mandatory technology in learning management systems and learners’ behavior. Therefore, it is a crucial factor that associates the relationship between EE and employees’ attitudes toward adopting the new system (Liu et al., 2023). Based on this identification, the following hypothesis is developed.

H5:

Faculties’ effort expectancy has a significant impact on their attitude toward adopting AI in higher education.

2.2. Applications of AI in higher education

The introduction of a digitalized learning approach changed the landscape of the higher education system (Khoza & Mpungose, 2022). A study by Carvalho et al. (2022) explored how society is going to foresee the future of education with a collaborative approach between learners, teachers, and AI. One of the primary applications of AI in higher education is to improve the learning experience for students (Ge & Hu, 2020). Additionally (Chang et al., 2022; Kelly et al., 2023), identified how the adoption of AI has changed the perception of society toward education. Meanwhile, studies have raised genuine concern about adopting AI in education by pointing out its impact on learners’ and users’ demographic, cultural, and behavioral issues (Chang et al., 2022). These studies strongly highlighted the relationship between AI adoption for society and its influence on users’ attitudes. Therefore, the following hypothesis developed.

H6:

Faculties’ adoption of AI for society has a significant influence on attitudes in applying AI in higher education.

2.3. Applications of AI and faculties work engagement

Artificial intelligence has the potential to revolutionize the way teachers engage with students and perform their roles in higher education. AI tools have been used in many institutions to engage in learning activities more productive way (Cui et al., 2019). The studies emphasized users’ attitudes toward adopting AI for personalized professional development, course design, grading and assessment, and student support (Franzoni et al., 2020; Rahimi & Tafazoli, 2022). Recent studies explored teachers’ attitudes and behavior in engaging AI-integrated CRM system and their digital competencies which enhance work engagement (Chatterjee et al., 2021; Ng et al., 2023). Further, Moreira-Fontán et al. (2019) explored the positive emotions and attitudes of academic staff members toward ICT-related aspects that boost their work engagement. Based on these findings the relationship between users’ attitudes and behaviors towards AI on work engagement is inevitable. Therefore, the following hypothesis was developed.

H7:

Attitude and behavior mediate work engagement in AI-adopted higher education.

Work engagements are closely associated with motivation and enthusiasm (De Simone et al., 2016). Highly engaged teachers are more likely to be open to accepting new innovations and digitalized teaching approaches (Scherer et al., 2019). Numerous studies identified that work engagement is highly associated with a higher level of adaptability and resilience in considering digital innovation in higher education (Al-Takhayneh et al., 2022; Antonietti et al., 2022). Similarly, a study quoted by AlAjmi (2022) notified the adoption of digital technology has enhanced teachers’ work engagement in higher education. All these observations strongly emphasize the association between work engagement and the adoption of advanced technology in higher education. Considering this broad view of previous research, the following hypothesis has developed.

H8:

Faculties’ work engagement significantly influences adopting AI in Higher education.

Users’ behavior in adopting new technology plays a vital role in forming interactions and collaboration with AI applications (Abumandour, 2022). Many studies identified teachers’ behavior plays a significant role in transforming higher education into an innovative digitalized mode (AL-Nuaimi et al., 2022; Müller & Leyer, 2023). Therefore, the literature provides sufficient evidence to roll out that user behavior plays a critical role in applying AI in higher education. Therefore, the following hypothesis was developed.

H9:

User behavior plays a significant influence in applying AI in higher education.

As per the existing literature, studies attempted to identify the role of AI and its potential to support teachers in higher education. However, existing literature is unable to explore the relationship between UTAUT with work engagement and the application of AI in higher education. Further, the current study also identified the potential research gaps in the intersection of artificial intelligence and its implication in higher education performance and the mediating role of behavior, attitude, and work engagement of users in applying AI.

The literature review summary in Table 1 reveals that various studies have investigated different facets of AI in higher education. Nevertheless, there is a gap in comprehensive research exploring the transformative potential of AI in higher education, particularly in understanding how these facets contribute to the integration of AI and the improvement of faculty engagement. The current research aims to fill this gap by focusing on these distinctive aspects.

Table 1. Summary of previous studies on AI in higher education

Objectives	Methods	Research Objects	Key Findings	References
Application of AI and Learning Environment	Surveys, Interviews	Faculty members	Teachers’ motivation to apply AI depends upon the problem presented by AI. Application is AI in HE depends upon its opportunities.	(Bucea-Manea-Țoniş et al., 2022; Chan, 2023)
Impact of AI on curriculum design and future direction	Experimental Designs	Higher Education Institutions	Identified AI mode of teaching methods to influence students’ learning through learning Management systems and curriculum design.	(Kuleto et al., 2023)
Reforming education in an era of AI	Comparative Analysis	Higher Ed Institutions	Explored how AI adoption resulted in institutional transformations, emphasizing improved accessibility, and learning outcomes.	(Qu et al., 2022; Xue et al., 2022; Zang et al., 2022)
Challenges in the Implementation of AI in Higher Education	Survey and Case Studies	Universities and Colleges	Identified challenges, ethical concerns, and barriers hindering effective AI integration in higher education settings.	(Chatterjee & Bhattacharjee, 2020b; Michel-Villarreal et al., 2023; Slimi & Carballido, 2023)
Faculty Role Transition	Qualitative Analysis	Teaching Professionals	Explored the evolving role of faculty in AI-driven learning environments, emphasizing a shift toward facilitation.	(Hwang et al., 2020; Naylor & Nyanjom, 2021)

Source: Literature Review; Scopus and Science Direct Database.

3. Conceptual framework and research model

Based on the concept of the “Technology Acceptance Model (TAM)” developed by Davis (1989), and “Unified Theory of Acceptance and Use of Technology (UTAUT)” developed by Venkatesh et al. (2003), the study proposed a theoretical framework that would illustrate the impact of facilitating condition, awareness, perceived risk, perceived expectancy, effort expectancy, adoption of AI in society on attitude and behavior which in turn towards work engagement and application of AI in higher education (Sohn & Kwon, 2020a). Both TAM and UTAUT theoretical models describe and predict the variables that influence user acceptance and use of information technology. TAM suggests that the users’ adaption and use of technology are influenced by users’ perceptions of its usefulness and ease of use (Dulle & Minishi-Majanja, 2011). These experiences in turn are influenced by various factors, such as previous experience, attitudes, and social norms. This model provides a useful framework for understanding the factors that influence the adaption and use of AI technologies in higher education (Popenici & Kerr, 2017). By identifying the key factors that influence adoption and use, TAM can help inform the design and implementation of AI technologies that are more likely to be adopted and used effectively by faculty and students. Similarly, UTAUT is an extension of the previous model “The Technology Acceptance Model” (TAM) and contains supplementary factors that influence user behavior. The detailed narration and description of variables have been depicted in Table 2.

Table 2. Narration and description of variables

Variables	Narration/Description
FC -Facilitating conditions	Facilitating conditions -refers to the extent to which an individual perceives that technical and related infrastructure are readily accessible and capable of supporting the utilization of AI in the system
AW-Awareness	Awareness about AI-based academic tools for routine academic activities.
PR – Perceived Risk	Perceived Risk refers to the belief that a user might experience an unfavorable outcome or loss while pursuing a particular outcome especially associated with a web-based technology like AI. It is closely linked with behavioral insecurities related to volatility associated with internet connections and functions.
PE- Performance Expectancy	Performance Expectancy refers to the relative advantage or usefulness of a new system like AI in the education arena.
EE – Effort Expectancy	Effort Expectancy refers to the simplicity or ease of use of a new system and is a significant predictor of Attitude.
A- Adoption	The adoption of AI for the improvement of society and making education cost-effective through AI-enabled learning is checked here.
AT- Attitude	To perform a target behavior positive or negative, feelings are exhibited by individuals. In this context, it refers to the attitude towards the use of new AI-based systems in the field of education.
BI - Behavioural	BI - Behavioural Intention is associated with a sense of assessing the strength of intention of an individual to perform a specific behavior.
WE- Work Engagement	Work engagement here refers to the way the classroom engagement of students is enhanced through AI-enabled tools. The faculty’s perception of improvement in the participation and performance of students through AI-facilitated pedagogy is investigated.
AI-HE	AI HE- Artificial intelligence in Higher Education – this variable refers to the way the teachers are using AI to create content for the classroom delivery, courseware, check plagiarism, and enhance the teaching experience using AI-enabled tools

UTAUT proposes four factors that significantly influence user acceptance and use of technology (Williams et al., 2009). Performance Expectancy is an individual who believes that using technology will enhance their job performance. The Effort Expectancy focused on the degree of ease associated with using technology (Sanusi, 2022). Similarly, Social Influence is the degree to which an individual perceives that important others believe they should use technology and finally facilitating conditions is the degree to which an individual believes that technical support and resources are available to support the use of technology (Mohamed Zabri et al., 2023). Based on these two theoretical frameworks, the study presented the conceptual framework in Figure 1 and proposed nine hypotheses. All these hypotheses will be tested with the support of empirical analysis.

Figure 1. Conceptual framework.

Source: Authors conceptual development.FC-facilitating condition; AW-awareness; PR-Perceived Risk; PE-Performance Expectancy; EE-Effort expectancy; AT-Attitude; BH-Behavior; WE-Work engagement; AAIH-Application of AI in higher Education

Figure 1 illustrates the study’s conceptual framework, depicting the relationships among independent, mediating, and dependent variables. The arrows in the figure indicate the associations, with an arrow pointing from the independent variable to the mediating variable, followed by arrows directed toward the two dependent variables. Table 2 provides a comprehensive overview of each variable outlined in this study, encompassing a total of 10 variables, including independent, mediating, and dependent variables.

4. Methods

The study has been conducted by clearly defining the research problem, which involves identifying the research questions and objectives. Since the objective of this research is to understand the impact of AI on higher education and investigate the adoption and implementation of AI in higher education, we developed a structured questionnaire to collect the data from respondents. A five-point Likert scale was used to measure the variables listed in Table 2. A survey-based primary investigation was conducted to collect data, employing multiple statements for each variable to comprehensively capture respondent opinions. In total 47 questions were included that can be classified as follows: Awareness (Q. no 1–4), Perceived Risk (Q. no 5–9), Performance Expectancy (Q. no 10–14), Effort Expectancy (Q. no 15–19), Facilitating conditions (Q. no 20–24), Attitudes (Mediating variables): (Q. no 25–29), Behaviour (Mediating variables) (Q. no 30–34), Adoption of AI for society (Q. no 35–38), Work engagement (Q. no 39–42), Application of AI in higher education (Q. no 43–47). The detailed questionnaire is attached in Annexure 1.

Figure 2 illustrates the research process relevant to this study. The initial segment of the figure highlights key sources for the concept, namely literature and theory adopted for the study. Subsequent elements of the figure detail the progression of how the outcomes were derived.

Figure 2. Research process.

Source: Developed by authors

4.1. Research design and data collection

The data has been collected from academic staff members from different designations, and specializations serving higher educational institutions and Universities that are delivering courses in hybrid modes from different parts of Asia. Based on the population of the study, 250 samples have been collected based on Cluster and Multi-stage sampling methods. The researcher distributed structured questionnaires via various social media (WhatsApp, Gmail, and LinkedIn). The detailed distribution of sampling is depicted in Figure 3. The collected data were coded by applying a statistical tool SPSS 29 and Smart PLS used to develop Structural equation modeling (SEM) to study the relationships between observed variables and latent variables. The outcome of SEM is used to test the fit of a proposed model to observed data, evaluate the theoretical framework, and refine the model (Figure 2).

Figure 3. Distribution of sampling collection.

Source: Developed by authors.

Figure 3 displays the amount of sampling collection distribution in each social media. The study collected sampling using the Google form link: https://docs.google.com/forms/d/e/1FAIpQLSfQl-ciiB-oem94ElF5vRTrQyF0FGYq2LPr1xgqtKPphSEN7A/viewform and distributed to all the sample using either WhatsApp, or Gmail, or LinkedIn, depending on their accessibility. Detailed allocation has been given in Figure 3.

4.2. Instrument’s reliability and validity

The instrument was developed considering (Davis, 1989; Venkatesh et al., 2003) TAM and UTAUT theory, which provides useful frameworks for understanding users’ acceptance and use of technology in higher education. A total of 10 independent and dependent variables were designed based on the review and each variable was represented with a set of questionnaires (Appendix 1). The researcher has developed a set of 45 statements/questions to measure various aspects of AI technology in the higher education sector. These questions cover topics such as familiarity with AI tools in higher education, customization of educational content, AI-powered chatbot technology, ease of use of AI technology, perceived risks of using AI technology, and the use of AI technology in handling admission procedures, impact of AI on attitude and behavior, adoption of AI on work engagement and application of AI tools in regular academic practices.

4.2.1. Factor loading

Factor loadings demonstrate the relationship of each item with the principal component, ranging from + 1 to − 1. Higher factor loadings, closer to + 1, indicate a stronger representation of the construct (De Roover et al., 2022). The findings show that all items exceeded 0.5, meeting the threshold criteria as per (Li et al., 2002). Items with loadings below 0.5 were subsequently eliminated.

4.2.2. Multicollinearity indicator

The study also administered a multicollinearity test which identified the consistency of the instruments. The measured variable Variance Inflation Factor (VIF) statistics is not above 5 which is the threshold as per (Miles, 2014). Since the value is as per the acceptable limit, it is considered that independent and dependent variables correlate with each other. The figures are depicted in Table 4.

4.2.3. Construction of validity

The study has administered Convergent Validity and Discriminant validity tests using PLS-SEM. The results indicate that the Average Variance Extracted (AVE) is greater than 5 and therefore, convergent validity is established according to (Olapade et al., 2023). Similarly, discriminant validity has been administered using the Fornell and Larcker criterion. The current values of the square root of AVE are greater in correlation with other variables. Therefore, as per Roemer et al. (2021), all the conditions were fulfilled. The figures are depicted in Table 4.

4.2.4. Reliability test

To measure the reliability of the instrument, the researcher administered Cronbach’s alpha test by considering a homogeneous sample of 30 and deleted some of the statements that were less reliable. The instruments were adapted from different studies and after the reliability test, it has been modified. The alpha value of the research notifies that all the values are above 0.700 which is above the threshold (Taber, 2018).

Table 3 presents the 10 variables along with the corresponding number of items within each construct. Additionally, the table displays Cronbach’s alpha values and sources from which the variables have been adapted.

Table 3. Constructs, measures, and reliability

Constructs	No. of items	Reliability (α)	Sources
Facilitating Condition	5	0.849	(Venkatesh et al., 2003)
Awareness	5	0.799	(Davis et al., 2023)
Perceived risk	4	0.711	(Davis et al., 2023)
Performance Expectancy	5	0.835	(Venkatesh et al., 2003)
Effort Expectancy	5	0.746	(Sohn & Kwon, 2020a)
Adoption of AI for Society	4	0.845	(Chatterjee & Bhattacharjee, 2020a)
Attitude	5	0.865	(Braganza et al., 2021)
Behavior	5	0.848	(Sohn & Kwon, 2020a)
Work engagement	4	0.799	(Braganza et al., 2021)
AI in higher education	5	0.845	(Chatterjee & Bhattacharjee, 2020a)

Source: Data Analysis

4.3. Participants and selection criteria

The sample for the study is geographically scattered and has been chosen from different parts of the Asian continents (India, Bahrain, Saudi Arabia, UAE, Oman, Qatar, and Sri Lanka). The researchers approached faculty members of various Universities and educational institutions that are either operating or currently operating in a hybrid mode of delivery. The selection of the sample has been conducted by administering cluster and multi-stage methods. Using Yamane (1967) formula, the existing populations spread across Asian countries, the sample was determined to be 250 respondents. The time frame for the sample ranged for a duration of 3 months commencing from January to March 2023. The questionnaire was administered to 300 respondents and with a response rate of 85% of the filled questionnaires were considered. Then by eliminating the lacunae the faulty and incomplete questionnaires were discarded and the final 250 questionnaires were used for further analysis. The sampling techniques and data collection framework are depicted in Figure 4.

Figure 4. Sampling technique.

Source: Developed by authors

Figure 4 elucidates the sample classification and its correlation with the population. The figure initially centers on the population, institutions covered, and the allocation of the sample across various counties, depicted with corresponding percentages.

4.4. Data analysis and interpretation

The data has been analyzed by applying various software like SPSS 29 and Smart PLS 29. The characteristics of the sample have been interpreted with descriptive statistics, and the t and p tests were administered to identify the significance of the variables by using SPSS. Researchers applied structural equation modeling through SPSS and Smart PLS to identify the relationship between different variables.

5. Results

5.1. Sample characteristics

The sample of the research was restricted to academic staff working in different parts of Asia with different designations and other characteristics. The sample consisted of 66 percent of male and 44 percent of women respondents. The survey respondents are distributed across various regions. However, the majority of participants in this survey come from higher education institutions in India, specifically those with a Quacquarelli Symonds (QS) Ranking and hybrid mode of delivery, accounting for 45% of the total respondents. On the other hand, Sri Lanka had the lowest representation, with only 5%, while the remainder hails from other GCC countries, including the UAE (17%), Bahrain (12%), Saudi Arabia (10%), Qatar (8%), and Oman (2%). For education, most of the respondents hold a master’s degree (50%), followed by a Doctor of Philosophy (48.33%) and a post-doc (1.66%). In age, the dominant respondents belong between 40 to 50 years (38.33%), followed by 30 to 40 years (31.66%), 23 to 30 years (18.33%), and 50 to 60 years (11.67%). In program specialization, most of the respondents are Commerce and Management (78.33%) graduates, engineering (10%), Humanities and social science (5% each), and medical (2%). In designation, most of the respondents are assistant professors (65%), followed by associate professors (25%), professors (8.33%), and teaching assistants (1.67%). In the mode of delivery, most of the respondents are in the regular classroom delivery (61.67%), followed by hybrid mode (23.33%), and distance education (15%).

5.2. Correlation and validity

Multicollinearity is a common issue in regression analysis, including PLS regression. It occurs when the predictor variables in a model are highly correlated with each other, making it difficult to distinguish the individual effects of each variable on the outcome variable. To address multicollinearity in PLS regression, the researcher administered principal component analysis (PCA). This method helped to reduce the correlation between predictor variables and improve the stability of the model. The threshold for acceptable VIF values is up to 5, others may set a lower threshold of 3.3 or even lower. Higher VIF values mean that the variance of the estimated regression coefficients is inflated, leading to less precise and less stable estimates (Table 4).

Table 4. Correlation and validity

	AD	AIHE	AT	AW	BH	EE	FC	PE	PR	WE	VIF	CR	AVE
AD	0.829										3.8	0.897	0.687
AIHE	0.716	0.784									3.7	0.887	0.614
AT	0.761	0.710	0.809								3.9	0.904	0.655
AW	0.439	0.420	0.558	0.744							3.4	0.861	0.553
BH	0.758	0.731	0.792	0.510	0.790						3.9	0.892	0.624
EE	0.542	0.611	0.628	0.561	0.533	0.859					4.0	0.849	0.738
FC	0.318	0.328	0.464	0.322	0.292	0.270	0.785				4.1	0.889	0.616
PE	0.672	0.641	0.808	0.545	0.703	0.609	0.369	0.818			3.8	0.889	0.668
PR	0.165	0.101	0.189	0.008	0.252	0.062	0.115	0.207	0.586		3.3	0.636	0.543
WE	0.800	0.690	0.739	0.432	0.661	0.575	0.310	0.698	0.115	0.790	3.9	0.867	0.624

Source: Data Analysis; Note: CR: Composite reliability; AVE: Average variance extracted.

The Average Variance (AV) of each construct is simply the square root of the AVE. Therefore, to establish discriminant validity, the researcher compared the AV of each construct to the correlation coefficients between that construct and other constructs in the model. Since the AV of a construct is greater than the correlation coefficients with other constructs, discriminant validity is established for that construct depicted in Table 4.

Based on the information provided in Table 5, the estimated model value of 0.08 for the standardized Root Mean Square Residual (SRMR) is higher than the commonly recommended threshold of 0.10, which suggests that the model does not fit the data well. However, it is still lower than the value for the saturated model, which indicates that the estimated model provides a better fit than the saturated model. The SEMR is 0.061 and 0.023 for structured and estimated models respectively, values are significant and as per the standard.

Table 5. Model fit

Mode	Structured Model	Estimated Model	Criteria
Chi-Square (X^2)	743.33	765.23	>0.01
SRMR	0.061	0.023	<0.08
NFI	0.935	0.912	>0.90

Source: Data Analysis

The NFI is a measure of the proportion of the χ² value of the null model that is accounted for by the proposed model. It ranges from 0 to 1, with values closer to 1 indicating a better fit of the proposed model. Specifically, the NFI is calculated as 1 - (χ² proposed model/χ² null model). An NFI value greater than 0.90 is generally considered to indicate a good fit for the proposed model. However, like the χ² value, the NFI is also sensitive to sample size and should be interpreted in conjunction with other fit indices. Since the NFI value is 0.935 for the structured model and 0.912 for the estimated model, the values are as per the required standard (Table 5).

5.3. Structural Equation Model (SEM)

Structural Equation Modeling (SEM) is a statistical method that allows researchers to test complex causal relationships between latent variables and observed variables. The study estimation is done using Smart PLS 4 and the outcome is depicted in Figure 5.

Figure 5. Structural model and path weight.

Source: Data Analysis

As depicted in Figure 5 and Table 6, a total of eight hypotheses are listed for analysis. Based on the results and analysis, all the hypotheses support the claim except hypothesis 5. This is due to the relationship between effort expectancy with attitude, being insignificant (p > 0.005, β = 0.057) respectively. Results show facility condition and attitude towards adaption of AI (Hypothesis 1) found to be significant (p < 0.001 and β = 0.83) and similarly in the case of awareness (AW) and Attitude (hypothesis 2) was also found to be significant with (p < 0.001 and β = 0.750) supporting the claim of faculties’ awareness towards AI towards their attitude to consider AI in higher education. Data depicts the significant relationship between hypothesis 3 perceived risk (PR) and attitude (p < 0.001 and β = 0.480), therefore hypothesis 3 is accepted. In the case of a relationship between perceived expectancy and attitude, (p < 0.001 and β = 0.420) is considered as significant, and thereby hypothesis 4 is accepted. Similar significant relationship outcomes were identified in the adoption of AI for society and attitude, (p < 0.001 and β = 0.480), so hypothesis 6 is also considered to support the claim. Since work engagement (WE) and the application of AI in higher education (AIHE) is a dependent factor, the study identified a significant relationship between attitude and work engagement (p < 0.001 and β = 0.704). Therefore, hypothesis 7 has been accepted. Similarly, work engagement also has a positive relationship with the application of AI in higher education (p < 0.001 and β = 0.427), therefore, hypothesis 8 is accepted. Finally, the behavior and application of AI in higher education are also considered to have a significant relationship (p < 0.001 and β = 0.427), so hypothesis 9 is accepted. The results indicate that FC, AW, PR, PE, and EE support the attitude to the tune of 79.4%, similarly, AT and AD support behavior to the tune of 68.3%. The mediating variable BH supports work engagement for the tune of 44.1% and BH and WE support the application of AI in higher education for the tune of 59.6%.

Table 6. Testing of hypothesis

Path	Hypothesis	β-value	Significance	Remarks
FC -> AT	H1	0.83	0.001	Supported
AW -> AT	H2	0.75	0.000	Supported
PR -> AT	H3	0.48	0.001	Supported
PE -> AT	H4	0.42	0.000	Supported
EE -> AT	H5	0.057	0.385	Not-Supported
AD -> AT	H6	0.480	0.000	Supported
AT-> WE	H7	0.704	0.000	Supported
WE->AIHE	H8	0.427	0.000	Supported
BH-> AIHE	H9	0.427	0.000	Supported

Source: Data Analysis

6. Discussion

Artificial intelligence has the potential to transform higher education in many ways, and hence it is mandatory to teach students to use AI-based algorithms. The effective integration of AI in higher education must be done with due diligence with thorough planning backed by ethical consideration (Bates et al., 2020). The findings of this study identified the application of AI in higher educational institutions by considering instances and experiences from different Asian countries. To explore the adaption of AI technology in higher education, we administered the UTAUT model and technology acceptance theories (Davis et al., 2023; Sohn & Kwon, 2020a; Venkatesh et al., 2003). The construction of the model consists of nine associations (hypotheses) between variables. The study has considered teaching staff in higher education as a respondent and identified their attitudes and behavior as mediating factors in adopting AI which supports the previous study by Chatterjee and Bhattacharjee (2020a) and identified the impact of these factors on the work engagement of faculty members which is district from all the previous research (Jiao et al., 2022; Sohn & Kwon, 2020b). The results have been interpreted with the help of different variables i.e., facilities and conditions, awareness, perceived risk, performance expectancy, effort expectancy, adoption of AI for society, attitude, and behavior. The results indicate that the faculty’s attitude and behavior have a significant impact on employee engagement and the application of AI in higher educational institutions.

One of the key questions raised in this study is the factors that influence the attitude of faculty members while adopting AI in higher education. The results identified that facility conditions create a positive impact on users adopting AI in their routine academic practices. This result correlates with previous study findings of Marks and Thomas (2022) identified that facility condition has a significant role in adopting new technology. The awareness of new technology among users emerged as a crucial variable in the current research. Awareness plays a vital role in the adoption of AI technology in higher education (Kour & Karim, 2020). The research findings indicate that the awareness of faculty members regarding the application of AI-based technology in their regular academic activities significantly influenced their work engagement. This, in turn, contributed to the integration of AI into higher education. This finding also supports the previous studies that notify institutions that provide training and professional development opportunities to help faculty members gain significant achievements in implementing new technology in their workplace (Sun et al., 2022). An essential contribution of awareness is its capacity to influence an individual’s attitude and approach toward unfamiliar aspects (Kang et al., 2023). The present study identified that awareness and attitude toward AI have a significant relationship which intern contributed positively to enhancing faculty work engagement. This outcome also supports the Marcinkowski et al. (2020) claim which identified that awareness of faculty members willing to learn about new technologies and how to apply them in their teaching and research has a significant influence on the application of AI. The results underscore the importance of implementing awareness, training, and learning programs within institutions to integrate new technology seamlessly into the organization. These results also support previous studies by Beerkens (2022) that identified that effective facility and awareness of new technology potentially transform higher education by providing more personalized learning experiences, improving student outcomes, and enhancing the efficiency of administrative processes.

The results explored perceived risk and expectation that hold a major role in embracing AI. The outcome of the study identified users’ apprehension on risk factors in adapting AI tools in administrative activities and student engagement. The findings explored how users are concerned about the learning process of AI systems in their routine tasks. However, results indicate that respondents are optimistic about enhancing their skills, collaborating more effectively, and improving student outcomes. These results correlate with the similar outcomes Teng et al. (2022) noted similar factors in applying AI in service sectors. Many educational institutions around the world are taking AI seriously and are investing in AI technologies to enhance teaching and learning, improve student outcomes, and increase efficiency (Sayed Al Mnhrawi & Alreshidi, 2022). Meanwhile, the study has identified insignificant associations between users’ effort expectance and attitudes towards adapting AI. These findings suggest that the efforts made by faculty members to learn AI-based technology do not have a significant impact on their attitude toward considering and implementing this system. However, previous research contradicts this finding and explores that effort expectancy has a significant role in enhancing the attitude of users in implementing new technology in their workplace (Al-Makhmari et al., 2022). It has reflected that between GCC and India, some of the institutions have established dedicated AI research centers, while others have incorporated AI into their curricula and instructional practices (Chaipongpati et al., 2022). Previous studies similarly proved the use of AI and other advanced technologies created a more flexible, responsive, and effective educational system that prepares students for the demands of the modern workforce (Gürdür Broo et al., 2022).

The finding further proves that performance expectancy has a significant positive impact on the attitude of faculty members toward the application of AI in higher education. Performance expectancy plays more importance in technological issues, and it improves individual faculties’ performance or outcomes (Silic & Back, 2017). These results indicate that the use of AI in higher education will lead to better teaching and learning outcomes, and they are more likely to have a positive attitude toward its adoption and use. It is essential to note that attitudes towards AI in higher education are affected by a series of circumstances, including perceptions of usefulness, ease of use, and social norms (AL-Ghamdi et al., 2022). Therefore, to enhance positive attitudes toward AI in higher education, organizations are required to provide adequate support, training, and resources to faculty members to facilitate them realistically incorporating AI into their teaching practices. It is notable that this empirical result is supported by the outcome of Chatterjee and Bhattacharjee (2020a), which state that both performance and effort expectancy are likely to have a positive attitude toward the adoption and use of AI in higher education. Their study opined that faculties are more likely to see the benefits of AI and view it as a valuable tool for enhancing teaching and learning outcomes, improving student engagement, and increasing efficiency.

Finally, the results indicate attitude and behavior mediate between work engagement and the application of AI in higher education. The results confirm that both direct and indirect effects of behavior and attitude have a significant impact on work engagement and the application of AI (Sharma et al., 2022; Moşteanu, N. R. 2022). Thereby the results express that the behavior of faculty members can also impact their willingness to adopt new technologies like AI in higher education, which can in turn influence their work engagement. The previous studies conducted by Kelly et al. (2022) reconfirm that individuals’ attitudes and behavior play a significant role in adapting AI in organisation. The outcome of this result indicates that faculty members who are open to change and willing to try new teaching and research methodologies are more likely to adopt AI and incorporate it into their work, which can lead to increased work engagement (Hooda et al., 2022). It is important for faculty members to be encouraged and recognized for their efforts to adopt AI in their work. When faculty members receive positive feedback and recognition for their innovative practices, it can help reinforce their engagement and motivation to continue exploring new ways of integrating AI into their work.

7. Conclusion

The current study identified the relationship of noticeable factors in the application of AI in higher education. The study noticed that factors like perceived risk, performance expectancy, and awareness significantly contribute to work engagement and application of AI in the higher education system through mediating variables namely, attitude and behavior. The study outcome indicates that the rapid advancements in AI technology have made it easier to implement AI solutions in various industries, including higher education. Based on the hypothesis testing, research notices the role of AI in improving personalized learning experiences, identifying at-risk students, and automating administrative tasks is evident. Finally, the outcome of the study emphasized applications of AI in higher education have the potential to improve faculty engagement and their attitude toward applying new technology in their routine teaching, learning, and assessment practices.

The findings of this study illuminate the transformative effects of swift advancements in AI technology, steering in an era of continuous incorporation across different industries, with a specific emphasis on higher education. The rigorous hypothesis testing piloted in the research highlights the critical and central role played by AI in advancing personalized learning experiences, adeptly identifying students at risk, and efficiently automating administrative tasks.

7.1. Implications

In recent years, there has been an increasing trend toward the adaptation of AI in higher education across Asia. Many universities and educational institutions have started to incorporate AI-powered tools into their teaching practices, such as intelligent tutoring systems, chatbots, and automated grading systems. However, the adaptation rate varies across different countries and institutions. This research suggests a multifaceted approach to higher education institutions involving various stakeholders, including legislators, educators, learners, and technology providers. Firstly, Policymakers should promote and create platforms by financial providing aid and infrastructural support system in universities, and educational institutions to embrace a culture of innovation and collaboration, where educators, students, and technology providers can work together to develop and implement AI-powered tools and solutions.

One of the key challenges in the implementation of AI is to tackle concerns regarding privacy and safety. As more students use online-learning platforms and share personal data, it’s important to ensure this information has been kept safe and confidential. It is thus the institution’s need to establish clear guidelines and criteria for the development and use of AI-powered tools in higher education. These guidelines should address issues such as data privacy, security, and ethical concerns. Since the outcome of the research identified that awareness and performance expectancy is one of significant importance to applying AI in educational institutions, it is essential that universities and institution must reserve budgets to provide training and support for educators to help them develop the skills and knowledge essential to effectively use AI-powered tools in their learning system. This may involve offering opportunities for professional development or partnering with technology providers to offer training programs.

It is imperative that policymakers prioritize the further development of infrastructure and technological advances in educational institutions to apply AI in higher education. To meet these needs, universities must prioritize developing intellectual capital and the resources to manage AI tools and technologies. The institution must be committed to developing systematic resource development programs by reserving a budget to manage AI-based software solutions, offering training and support for faculty and staff, and creating partnerships with industry-leading companies in AI research and development. Similarly, by making a systematic market survey, institutions must develop AI-based curricula and courses to prepare students for future jobs and provide them with the skills needed to succeed in an AI-driven world.

8. Limitations and suggestions for future research

The research identified that AI has the potential and scope to revolutionize the higher education system and improve faculty work engagement, however, there are still certain limitations that must be addressed. Since many institutions are skeptical about the effectiveness and implementation of AI, the adaption and acceptance of AI in higher education are still limited. This has led to concerns about the cost and complexity of implementing AI solutions. Furthermore, the application of AI in higher education raises concerns about privacy and security. For example, if AI assessment tools are applied to track student performance, there is the possibility of revealing personal information about students that they may not want to share. Furthermore, if AI is administered to make decisions regarding faculty work engagement, it could potentially lead to bias and discrimination. Therefore, we suggest future research possibly focus on the development of ethical guidelines for the application of AI in higher education. These guidelines must address privacy, security, and bias concerns and provide a framework for the effective use of AI in higher education. In conclusion, while the application of AI in higher education and faculty work engagement holds great promise, there are still significant challenges and limitations that need to be addressed. Since the application of AI in higher education and its impact on faculty work engagement is in its nascency stage, a lot more challenges and limitations need to be addressed. Further research would be required in order to develop ethical guidelines, especially concerning the privacy of the stakeholder, and also on various faculty engagement program that not only foster collaboration, but also leads to a partnership between academia and industry.

Acknowledgments

I would like to thank Kingdom University Bahrain for supporting and granting funds.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/2331186X.2023.2293431

Additional information

Funding

The work was supported by the Kingdom University Bahrain [2023-17].

Notes on contributors

Habeeb Ur Rahiman

Habeeb Ur Rahman is an innovative academician with over a decade of expertise in higher education. He holds a Ph.D. in Management and an MBA from VTU India. He has achieved the prestigious ‘Fellow’ status from Advance HE, and Manager from CMI from the UK. Dr. Habeeb has received certifications in FinTech from the University of Cambridge, Artificial Intelligence from MIT, Transactional Analysis from The Berne Institute, and ISO lead auditor. He is an active researcher in the field of social science and humanities, with publications in reputable journals indexed under Scopus, ABDC, and WoS. As an Asst. Professor and Head of the SDU at Kingdom University, Bahrain, he combines his deep understanding of behavioral sciences with his role as a skilled freelance trainer in self-development, behavioral diagnostics, transactional analysis, and emotional intelligence.

Rashmi Kodikal

Rashmi Kodikal is an experienced and innovative professor with 21 years of teaching expertise at the postgraduate level. She has contributed significantly to research, presenting 112 papers at international and national conferences and publishing 54 articles in renowned journals indexed under Scopus, Web of Science, and ABDC. Her research focuses on organizational behavior, service quality, and sustainability. Dr. Kodikal has served on various academic boards and panels and is currently the Chief Editor of two national-level management journals. She is also a skilled freelance trainer in behavioral sciences, conducting programs on self-development, behavioral diagnostic tools, transactional analysis, and emotional intelligence. Currently, she serves as a Professor of Management Studies at Graphic Era Deemed to be a University in Dehradun, India.

Appendix 1:

Survey Questionnaire

Awareness

• 1. I am familiar with data transformation and artificial intelligence-based academic tools.

• 2. Artificial Intelligence tools are highly useful to prepare educational content and materials.

• 3. AI-based technology like a chatbot quickly provides information and answers queries about academic affairs.

• 4. I am aware of the application of AI-based technology in routine academic activities.

Perceived Risk

• 5. I am aware of ethical aspects related to AI applications.

• 6. I believe AI-powered educational content is not always correct.

• 7. The application of AI for admission purposes is confusing.

• 8. I shall not prefer to use AI applications for administrative purposes.

• 9. The use of AI technology for answering student’s queries is risky.

Performance Expectancy

• 10. It will be hard to develop a perfect AI application catering to the needs of administration in Higher education.

• 11. AI-powered learning activities will enhance the efficiency of the higher education system.

• 12. Educational content prepared by AI technology is useful.

• 13. Using AI-powered chatbot technology I can get an accurate answer.

• 14. Smart educational content can be prepared using AI technology.

Effort Expectancy

• 15. AI technology is not easy to learn.

• 16. I need to put a lot of effort into learning AI technology.

• 17. If I know basic AI technology, I can easily learn AI-based applications.

• 18. I can have my query answered quickly using AI chatbot technology.

• 19. Individualized content can be prepared using AI technology.

Facilitating conditions

• 20. My institute has all the necessary resources to use AI technology for smart.

• 21. I have all the required resources to develop AI-based smart content.

• 22. My institute sponsors any AI-related learning opportunity.

• 23. All the classrooms of my institute are equipped with the necessary devices for using AI technology for teaching purposes.

• 24. My institute encourages its staff to use modern technology.

Attitude

• 25. I can learn AI technology quickly.

• 26. AI technology is useful for teaching-learning activities.

• 27. Using AI technology for query answering is a good idea.

• 28. People should learn AI technology for the future need of the higher education sector.

• 29. AI technology can cater to individual needs more accurately.

Behavior

• 30. I believe AI technology is very easy to learn for beginners.

• 31. I am willing to use AI technology for developing smart content.

• 32. I believe AI technology could be used for answering student’s queries.

• 33. I shall recommend all the stakeholders in higher education explore AI.

• 34. I intend to use AI technology for teaching-learning purposes in the next couple of years.

Adoption of AI for Society

• 35. The application of AI in higher education is good for society.

• 36. The application of AI in higher education will make education more interactive.

• 37. The application of AI in higher education will make it cost-effective.

• 38. The application of AI in higher education will make teaching-learning activity more interesting.

Work Engagement

• 39. AI technology made my learning and teaching experience more interactive and interesting.

• 40. The adoption of AI technology and tools fostered my classroom engagement.

• 41. AI technology developed my performance and engagement in research activities.

• 42. AI technology and resources are enhanced by participation and performance in professional and personal development activities.

Application of AI in HE

• 43. I apply AI technology to create teaching material and content development.

• 44. I apply AI tools to review homework, tests, and other written assignments, monitor student achievement, and provide feedback.

• 45. I apply AI tools to detect plagiarism in student papers and courses works.

• 46. The application of AI in my higher education academic journey is cost-effective.

• 47. I am using AI technologies and tools in my teaching and learning activities.