Since the mid-20th century, universities have undergone significant changes due to increased digitization. One of the most significant advances has been the emergence of artificial intelligence (AI). In 1956, John McCarthy’s team at the Dortmund Symposium pioneered the early development of AI, establishing a new science aimed at developing systems that mimic human cognitive abilities ( Norvig and Russell, 2010). Since then, AI has transcended from a branch of computer science to become a powerful tool, boosting performance in diverse fields such as healthcare, transportation, and entertainment. These sectors have benefited from its ability to automate most academic and administrative tasks ( Saxena and Doleck, 2023). As research has advanced, AI has been applied to techniques that mimic human reasoning, such as self-learning, prediction, and decision-making ( Abrokwah and Larbi-Oko, 2024; Zakharov, 2021). Recent research suggests that the effectiveness of these technologies in education depends crucially on whether educational institutions can enhance students’ competence in areas such as digital literacy and data literacy and cultivate the technical skills necessary to effectively use these tools ( Wang, Sun, and Chen, 2023). Research indicates that artificial intelligence (AI) extends beyond machine learning or data processing to encompass artificial neural networks and language processors capable of adapting to new data and predicting outcomes ( Baker and Smith, 2019; Popenici and Kerr, 2017). ( Luckin et al., 2016) categorize AI applications in education into three main areas: intelligent tutoring systems, collaborative tools, and virtual reality technology. These tools help design personalized learning paths for students, enhance interactivity, and provide more diverse learning experiences ( Salmon, 2000; Jonassen et al., 1995). In this context, it can be said that AI constitutes an integral component of higher education systems at three different levels: at the student level, through personalized educational support; at the faculty level, through tools to improve teaching; and at the institutional level, through systems that support management and strategic decision-making ( Joseph, 1999). Thus, we propose the following: H1:

Artificial intelligence has a positive and statistically significant impact on promoting digital innovation at the University of Ninevah.

Digital innovation enables the use of digital technologies in the development of services and products, enhancing institutional integrity and value ( Nambisan et al., 2017). The concept of digital innovation is not merely an expansion of traditional innovation; rather, it represents a qualitative shift in the creation of new, more innovative, and creative practical solutions. These solutions rely on digital platforms and smart application practices that facilitate collaboration between all stakeholders and enable them to access and enter markets in innovative, unconventional ways. The importance of digital innovation is also evident in the integration of physical and virtual systems, resulting in hybrid products and services that combine the physical and the digital, such as smart devices and interactive applications. However, these transformations have a significant impact on traditional leadership, necessitating a leadership style capable of handling the complexity and rapid developments resulting from digital transformations ( Svahn et al., 2017). In Arab and Iraqi societies, digital innovation is still in its infancy and faces obstacles related to a lack of infrastructure and a lack of digital literacy among academic staff. This necessitates the development of long-term plans that create a supportive environment for innovation and integrate human, technical, and organizational resources ( Ahmed et al., 2022). It is recognized that the success of digital innovation depends on the ability of educational institutions to establish sustainable and appropriate governance frameworks for the educational process, with a focus on improving the quality of their academic outcomes, expanding the scope of blended learning, and developing digital curricula that align with global trends ( El Said, 2021). Thus, we propose the following: H2:

Artificial intelligence has a positive and statistically significant impact on supporting digital transformation at the University of Ninevah.

Digital transformation plays a significant role in integrating digital technology into various aspects of business, including the education sector. The goal is to improve customer service and enhance operational efficiency. Furthermore, it includes redesigning a company’s or organization’s structure, changing its culture, and developing its leadership style to keep pace with new changes ( Westerman et al., 2014). ( Liu et al., 2023) explained that digital transformation cannot be limited to developing technological capabilities, but can be expanded to include moving from the traditional approach to an integrated business model based on digital technology to create new products and services. Digital transformation encourages the creation of new revenue streams by enhancing the value of services provided to the public, and education services are certainly one of these. Meanwhile, ( Vial, 2019) noted that “digital transformation is a complex process that requires continuous interaction between technological resources and organizational capabilities,” leading to significant and clear changes in the business structure ( Matt et al., 2015). Rather, it depends primarily on the ability of administrators and leaders to build a distinct digital culture within the organization and educate employees on the importance of adapting to the requirements of the digital age. This also applies to educational institutions such as universities and other institutions operating in this field. Digital transformation has become an important and necessary aspect of improving the quality of education and academic knowledge management. Universities that have implemented digital transformation and pursued a strategy of adopting digital technology have been able to establish a broad base for an educational system based on e-learning and have acquired significant capacity to handle big data. It has also been able to provide more comprehensive, flexible, and modern educational services, enhancing its capacity for innovation in various forms and across diverse disciplines, and ensuring the sustainability of its institutional performance over the long term ( Benavides et al., 2020). Study ( Hess et al., 2020) clarified the differences between “digital transformation” and simple “digitization.” Simple digitization refers to the conversion of paper data into digital form, while “digital transformation” refers to a strategic shift aimed at creating new value, under clear and advanced leadership with effective management capable of absorbing, leveraging, and developing change to achieve comprehensive, continuous, and sustainable development ( Reis et al., 2018). ( Verhoef et al., 2021) explained that digital transformation is now considered a prerequisite for organizational success, rather than a secondary option, as it contributes to improving customer experience, increasing organizational agility, and enhancing the ability to innovate. ( George and Paul, 2020) also view digital transformation as a complex process that relies on the continuous interaction between modern technology, human resources, and the organization. Practical experience shows that organizations that implemented integrated digital transformation strategies were better able to deal with crises, such as the COVID-19 pandemic, which accelerated digital transformation across various sectors, especially education Figure 1 shows the proposed relationships in this study, Thus, we propose the following: H3:

Digital transformation has a positive and statistically significant impact on promoting digital innovation at the University of Ninevah.

Participants received an email explaining the study and inviting them to participate. Those who agreed clicked a link in the consent letter attached to the invitation, which redirected them to the survey. The email invitation was received by 581 faculty members currently working at the university, and a total of 206 faculty members responded (35% response rate). The sample comprised faculty members representing various academic and research ranks from various departments and universities, including faculty members working in residence at the university.

Ethical approval

This study was conducted at the University of Ninevah in Iraq. According to the institutional practices at the time of data collection, no formal ethics committee review process or approval number was available for studies involving academic staff in non-invasive survey-based research. However, the study adhered to the principles of the Declaration of Helsinki, and participation was voluntary, anonymous, and confidential.

Informed consent

Informed consent was obtained verbally from all participating university faculty members. Verbal consent was chosen because written consent is not required in non-invasive survey studies conducted within Iraqi universities, and because participants preferred fast and anonymous participation. All participants were informed about the purpose of the study, confidentiality of responses, and their right to withdraw at any time.

The demographic characteristics of the study sample are presented in Table 1, which shows that 55.8% (n = 115) of the respondents were men (compared to n = 91 women). The results also indicate that 62.1% of the respondents were teaching assistants, lecturers 26.6% were assistant professors, and only 9.4% were full professors 1.9%. The number of respondents under the age of 30 was 12, equivalent to 5.8% of the sample group of 206. The total number of respondents aged 30-39 was 147, equivalent to 71.4%, while those aged 40-49 were 29, equivalent to 14.1%. Those aged 50-59 were 17, equivalent to 8.2%, while those aged 60 and over were 1, equivalent to 0.4%. All university faculties participated in the survey, with 22.1% of respondents employed at the Faculty of Engineering, 16.5% at the Faculty of Information Technology, 17.5% at the Faculty of Law, 12.1% at the Faculty of Medicine, and 12.6% at the Faculty of Nursing (9.2%), the Faculty of Pharmacy (17%), and the University Presidium (15%).

All scales are composite. Each scale is a previously published and validated composite scale. Unless otherwise specified, the means of the items in each scale were:

Each scale is a composite measure, each of which has been previously published and validated unless otherwise specified. The means of the items in each scale were: Perceptions of Artificial Intelligence (AI), which was measured as an independent variable, composed of three dimensions: general perceptions of AI, current use of AI, and future impact of AI. Each of the three dimensions above consists of four items measured using a five-point (Likert) scale (1 = strongly disagree, 5 = strongly agree). Examples of items include “AI contributes to improving the quality of higher education,” “AI enhances interaction between students and faculty members,” and “I benefit from smart learning platforms that rely on AI to evaluate my academic performance.” The Cronbach’s alpha coefficient for this scale was 7.46, indicating good internal consistency. Digital transformation (DT) was also measured as an intermediate variable (M) consisting of three dimensions, each of which consists of four items: the institution’s digital strategy, the implementation of digital transformation in the learning process, and the university’s readiness for digital transformation. Examples of this are: “The university provides the necessary infrastructure for digital transformation,” “Digital transformation helps in monitoring academic progress better,” and “Universities that implement digital transformation achieve better performance.” A five-point Likert scale was used (1 = strongly disagree, 5 = strongly agree). The Cronbach’s alpha coefficient for this scale was (0.773), indicating good internal consistency. Digital innovation as a dependent variable (DI) consists of three dimensions: (technological development in education, the use of digital innovation in the learning process, and the future impact of digital innovation). Each dimension consists of four items. Examples of this include: “The university encourages digital innovation within its learning strategy,” “Provides technical support to users of digital systems,” and “The university provides the infrastructure.” Necessary for digital transformation. A five-point Likert scale was used (1 = strongly disagree, 5 = strongly agree). Cronbach’s alpha coefficient was (.797), indicating very good internal consistency.

In our current research, we used structural equation modeling (SEM) using AMOS software. For our hypotheses, we presented four main hypotheses: one explains the direct influence relationship between AI and digital innovation, and the other the indirect influence relationship between AI and digital innovation with the mediating variable, digital transformation. Before delving into the details of testing these hypotheses, it is necessary to properly test the significance of the hypothetical model and present it in its general form. It is necessary to test the relationships between the independent and dependent variables independently, between the independent and mediating variables, and between the mediating and dependent variables. We must ensure that these tests include significant influence relationships so that the hypothetical model can be comprehensively tested as a single package. Table 2 also presents the descriptive statistics, reliability coefficients, and correlations.

Figure 2 illustrates the direct positive impact of artificial intelligence on digital innovation at a statistical significance level of 0.05. This hypothesis was met, as evidenced by the fit values that were within acceptable levels. The Goodness of Fit (GFI) value was 0.948, and the AGFI (Adjusted Goodness of Fit) value was 0.937, which is above the acceptable levels (above 90%). The Root Mean Square Residual (RMSR) also reached 0.064, which is lower than the standard rate for model acceptance. According to ( Bentler, 1990), the RMR should not exceed 0.080, indicating the quality of the direct impact of artificial intelligence on digital innovation. This enables us to accept the first hypothesis, which confirms the existence of this direct impact. Figure 2, the direct positive impact of artificial intelligence on digital innovation. Table 3 shows the impact values of artificial intelligence on digital innovation in detail. The regression values were positive, amounting to 1.630, which is a direct, positive impact relationship, meaning that the more the use of artificial intelligence tools increases, the more digital innovation will increase, which confirms this positive impact. The confidence level values were also positive, indicating the significance of the impact, which enables us to accept the first hypothesis. The significance of the P value was less than 0.05, indicating the acceptance of the first hypothesis.

One of the requirements for proving the relationships in the hypothetical diagram is the second relationship between the independent variable (AI) and the mediating variable (digital transformation). This must have a direct, positive, and unidirectional impact. This was achieved through Figure 3, which illustrates the impact of artificial intelligence on digital transformation, as measured by the matching values, which were within acceptable levels. The GFI (Goodness of Fit) value was 0.902, and the AGFI (Adjusted Goodness of Fit) value was 0.880, which are within acceptable levels (above 90%, except for the AGF value). The RMR index was slightly higher than the acceptable level of 0.082, which is greater than 0.080. This does not represent a strong impact on the quality of the model, given that the remaining indicators are within acceptable levels, indicating the quality of the direct impact of artificial intelligence on digital transformation. To test the second hypothesis in detail, we note the positive impact relationship between artificial intelligence and direct digital transformation, as indicated by the impact value of 1.204, and at the positive confidence level within the highest level of 2.334 and the lowest level of 0.782, both levels are positive, indicating the significance of the model as shown in Table 4. Through the significance value of P, we can accept the second hypothesis, which states that there is a significant impact of artificial intelligence on the mediating variable, direct digital transformation. This is a second proof of the acceptance of the hypothetical model and the significance of the impact between the independent and dependent variables.

Figure 4 shows the matching values between the mediating variable (digital transformation) and the dependent variable (digital innovation). The impact relationship was positively proportional; the greater the digital transformation, the greater the digital innovation, as indicated by the impact value of 0.892 in Table 5. The Root Mean Square Residual (RMSR) was 0.071, which is lower than the standard rate for model acceptance and should not exceed 0.080. The GFI and AGFI values reached 0.92 and 91%, respectively, indicating the quality of the hypothetical scheme in terms of the impact of the mediating variable (digital transformation) on digital innovation. The more data is transformed from paper to digital, the greater the impact on digital innovation in the organization under study. Regarding the indicators of testing the H3, we note that it has been achieved at the level of the current research. Through the value of the positive and direct effect of 0.892, we note that the third hypothesis (the impact of digital transformation on digital innovation) has been achieved. The confidence levels were significant, as they were in the correct direction, which is a one-way positive direction indicating the significance of the effect. The highest level of confidence reached 1.458 and the lowest 0.566, both of which were in one direction positive indicating the significance of the effect. This is confirmed by the p value, which was significant and reached 0.006, which enables us to accept the third hypothesis, which states: “There is a significant impact of digital transformation on digital innovation at the significance level for human research, which is 0.05”.

After testing the components of the hypothetical scheme presented by the current research, we can now, through Table 6 and Figure 5, test the fourth hypothesis, which expresses the research objective, problem, and basic idea: “Does artificial intelligence impact digital innovation, with the presence of the mediating variable, namely digital transformation?” Figure 5 shows that we can present the hypothetical diagram as a single package and it can be used after a long period of time so that it gives the same results because we tested its parts and the influence relationships it presented and they were all significant and with acceptable confidence levels. From Figure 5, it is clear that the value of the square root of the mean of the residuals (RMR) was within the acceptable rate, which amounted to 0.070, indicating the model’s ability to represent the influence relationship between the variables. The model is also characterized by quality through the GFI and AGFI values, which amounted to 0.917 and 0.905, respectively, indicating the quality of the model and the absence of significant differences between the hypothetical model presented by the current research in the methodology and the applied model on which the opinions of the respondents in the researched organization were surveyed. Table 6 shows the detailed impact relationship and the degree of significance of the direct and indirect impact, and whether the role of the mediating variable has a complete impact, a partial impact, or no impact of the mediating variable. From Table 6, we note that the value of the direct impact was 0.865, which represents a direct relationship, i.e., the more the use of artificial intelligence increases, the more digital innovation increases. This is a logical relationship, as artificial intelligence provides many solutions in various academic and research fields. The confidence levels at the 95% level were at the lowest level of 0.402 and the highest level of 1.766, which is positive in one direction, indicating the significance of the direct impact, as indicated by the p-value, which reached 0.005, which is less than the standard significance level in human studies, which is 0.05. The essence of the model and the title of the current research can be tested in the relationship between the independent variable, artificial intelligence, and the dependent variable, digital innovation, in the presence of the mediating variable, digital transformation. This was achieved through the data in Table 6, as the value of the positive indirect effect reached 0.554, which is a positive effect that explains a logical relationship despite the entry of the mediating variable. In other words, there is an effect of artificial intelligence on digital innovation, but on the condition of the presence of digital transformation. The more artificial intelligence and digital transformation increase, the more digital innovation increases in the organization under study. This value also indicates that the mediating variable has a partial effect, as the effect is distributed between the independent variable and the mediating variable to affect together the dependent variable, digital innovation, as indicated by the lowest level of 1.208 and the highest level of 2.71 of the confidence levels and as indicated by the P value of 0.002, which is greater than the standard level in human studies of 0.05. This also indicates the significance of the indirect effect, in other words, the distribution of the effect between the independent variable and the mediating variable in the dependent variable. In other words, in order to achieve digital innovation, the effect must be distributed between artificial intelligence and digital transformation in order for that innovation to be achieved, which enables us to accept the fourth hypothesis, which states, “There is an effect of artificial intelligence on digital innovation with the presence of digital innovation in the educational organization at a significance level greater than or equal to 0.05”.