Explaining the effect of artificial intelligence on the technology acceptance model in media: a cloud computing approach

Abstract

Purpose

The purpose of this paper is to explain the effect of the technology acceptance model in the media environment by using the mediating role of artificial intelligence and the cloud computing approach.

Design/methodology/approach

After reviewing the theoretical foundations, a conceptual model framework and research hypotheses were formed. The statistical population of the study included managers, deputies and experts from the National Iranian Oil Company, and a statistical sample of 368 people was selected by simple random sampling.

Findings

The results of structural equation modelling with PLS 2.0 software show a positive and significant effect on the artificial intelligence variable in the technology acceptance model with the cloud approach. Artificial intelligence has opened a new space in the digital world, especially in the media, so that its profound impact is quite evident and has affected people’s lives.

Originality/value

The acceleration of various technologies has severely challenged the approach of organizations, especially the media. The media environment with word of the technologies of the Industry 4.0, especially cloud computing technology, has changed the ways of accessing and using products and services.

Citation

Tavakoli, S.S., Mozaffari, A., Danaei, A. and Rashidi, E. (2023), "Explaining the effect of artificial intelligence on the technology acceptance model in media: a cloud computing approach", The Electronic Library, Vol. 41 No. 1, pp. 1-29. https://doi.org/10.1108/EL-04-2022-0094

Introduction

Reviewing the innovative periods in mankind’s history surprisingly talks about the desire of humans to move towards technological transformation (Na et al., 2022). Since Industry 1.0, mankind has always witnessed drastic technological changes referred to as the second and third “waves” (Morrar et al., 2017; Xu et al., 2018). We are currently witnessing Industry 4.0. Since the introduction of Industry 4.0 in January 2016, the topic of technological transformation led by artificial intelligence (AI) has been predicted to be very close (Skilton and Hovsepian, 2018).

AI in the field of technology offers organizations the opportunity to improve their performance (Borges et al., 2021). Research regarding AI has focused on how to reproduce aspects of human intelligence, including the ability to communicate, within the machine (Frankish and Ramsey, 2014; Guzman and Lewis, 2020). Today, this gulf between AI and communication research is narrowing, bridged by AI technologies designed to function as communicators. Recent advances in AI have led to more powerful and consequential AI technologies being integrated across daily life (Campolo et al., 2017). According to the Pew Research Center, individuals routinely chat with Amazon’s Alexa, Apple’s Siri and other digital assistants (Olmstead, 2017), with people’s interactions with smart devices expected to grow along with the emerging Internet of Things (Rainie and Anderson, 2017). Within the industry, media providers, such as the Associated Press, are using AI-enabled technologies in the production and distribution of news (Marconi et al., 2017). There is still limited knowledge regarding the integration of this technology in management practices, especially in cloud ERP systems (Yathiraju, 2022).

Electronic learning is a new system in which different types of technologies and tools are integrated together and play a major role in creating e-learning. E-learning systems often require a lot of hardware and software resources, which, nowadays, cloud computing technology has changed the development and availability of much software. There are many educational institutes that cannot afford such investments and cloud computing seems to be the best outcome for them. Using a cloud computing e-learning system has its own properties and requires a specific approach. Over the years of research by researchers, various terminology has been applied to the concepts of e-learning, media and technology: computer, bandwidth, multimedia, interactive, hyperlinked, collaborative, distance learning, simulated situation, hyperlearning, media-rich, more formal, structured, broadband, intelligent and usable (Kumar Basak et al., 2018; Laouris and Eteokleous, 2005; Sharma and Kitchens, 2004; Soualah-Alila et al., 2013; Traxler, 2007).

Various organizations, both in public and private sectors around the world, have been using cloud computing for some time, and more and more organizations are moving towards infrastructure and cloud technologies (Yaghubi et al., 2016). Cloud computing is recognized as an important area in information technology innovation, and massive investments have been made in it (Armbrust et al., 2010). Cloud computing technology is one of the most important topics in the field of information systems development, so cloud computing is considered a new paradigm for hosting and providing services on the internet (Avram, 2014; Lian et al., 2013). Cloud computing is a set of easy-to-use virtualized resources that can be dynamically modified to meet customer demand (Yektaei and Ranjbar Noshary, 2016). Cloud computing will certainly change the structure and nature of information systems of organizations in general, and, in particular, it will change the development of telecommunications. Nevertheless, before using cloud computing, organizations should seriously evaluate and choose the appropriate technological model for their organization (Kuo, 2011; Osei Yeboah and Kofi Asare, 2014).

Alhanatleh and Akkaya (2020) declared that cloud computing can be a new horizon whereby technological resources for computing (i.e. processing, memory and storage) are stored in a place other than the user’s physical location. Cloud computing is a model that enables easy access from anywhere and on-demand to a repository of configurable computing resources, such as networks, servers, repositories, applications and services. They can be provided or delivered quickly without the need for the intervention of a service provider (Avram, 2014). According to the National Institute of Standards and Technology, cloud computing has four key features: on-demand self-service, broad network access, rapid elasticity and measured service. The goal of some researchers is to study the concept of cloud computing by studying only one type of it, as an example of infrastructure as a service (Bhardwaj et al., 2010). Another group of researchers studies cloud computing on a specific system, such as the human resource planning system (Saeed et al., 2011), presented by providers or e-learning programmes.

Another study was conducted by Low et al. (2011), which examined the acceptance factors of cloud computing. They postulated the impact of eight factors on the adoption of cloud computing in high-tech industries in Taiwan. They were trying to identify the factors that led to the acceptance or non-acceptance of the cloud by the owners of these industries. Surveys have shown that a comparative advantage hurts cloud acceptance, as well as the support of senior management, organization size, competitive pressure and a business partner have had a significant positive effect on cloud propagation and acceptance. Also, in this study, compatibility and complexity in cloud acceptance had no significant effect (Low et al., 2011). Important variables influencing the use of cloud computing by researchers have also been presented over the years: organizational factors (information intensity, employee knowledge) (AbuKhousa et al., 2012; Karahanna et al., 2002; Premkumar and Roberts, 1999; Tornatzky and Fleischer, 1990; Yusof et al., 2008); human factors (decision-makers’ innovation, decision-makers’ knowledge, external support and competitive pressure) (AbuKhousa et al., 2012; Lian et al., 2013; Lin et al., 2012; Lin and Yen, 2011; Yusof et al., 2008); environmental factors (external support, environmental infrastructure and competitive pressure) (Chang et al., 2007; Lian et al., 2013; Thong, 1999; Tornatzky and Fleischer, 1990); technological factors (comparative advantage, complexity, testability, compatibility, technological infrastructure, security and privacy) (Chang et al., 2007; Gupta et al., 2013; Kuo, 2011; Lian et al., 2013; Lin and Yen, 2011; Premkumar and Roberts, 1999; Rogers, 2003; Sultan and Sultan, 2012; Tornatzky and Klein, 1982; Yusof et al., 2008). Gupta et al. (2013) in their research considered various factors effective, such as cost reduction, ease of use and persuasiveness, reliability, collaboration and sharing, security and privacy, in the use of cloud computing by small and medium enterprises. Bento et al. (2015), in their research on the motivation of organizations to use cloud human resource planning, included in their factors: faster access to new capabilities, increased revenue by offering new products, use of better resources and reduced costs; they also identified decentralization and reported barriers to cloud access: security, regulation, reliability, capability and maturity of cloud services, in addition to government regulations. In their research, Wang et al. (2015) addressed the opportunities and challenges that IT executives face in moving to cloud computing, such as data privacy, security, IT governance and local regulation when moving to the cloud.

Therefore, the necessity of using a technology acceptance model (TAM) in the organization with a cloud approach seems necessary. Importantly, the adoption of technology in this regard has been the subject of much research. Hence, different theories have presented new approaches to the acceptance and use of information technology in the individual and organizational dimensions. Among the studies conducted in the group of individual acceptance models, Venkatesh et al. (2003) introduced the following eight models as the most important TAMs that are the basis of other models: reasoned action theory, TAM and motivational model, theory of planned behaviour (TPB), combined TAM and planned behaviour, personal computer use model, innovation dissemination theory (IDT) and social cognition theory. In another section, the most important models of technology acceptance are presented as follows: theory of reasoned action (TRA) (Fishbein and Ajzen, 1975); IDT (Rogers, 1983); social cognitive theory (SCT) (Bandura, 1986); TAM1, the basic TAM (Davis, 1989); TPB (Ajzen, 1991); technology acceptance model 2 (TAM2) (Venkatesh and Davis, 2000); and unified theory of acceptance and use of technology (Venkatesh et al., 2003).

Table 1 compares some important indicators from the perspective of TAMs.

Examining these models, it can be concluded that they all follow the following conceptual structure (Figure 1).

In recent years, social networks have had a significant impact on people’s interactions. More and more users are using social media for entertainment, web browsing, news awareness or just to spend time (Sangari et al., 2020). Social media in the workplace, for example, provides employees with new ways to interact with their customers and co-workers and to ask questions and share information (Leftheriotis and Giannakos, 2014). Organizations can reap significant benefits through the use of social media. One of these benefits is the creation of an intelligent knowledge management system in the organization. Interactions between employees and the creation of group discussions cause the exchange of useful information between employees and enable them to improve their performance (Nisar et al., 2019). Organizations can use social media to save money, engage with customers and build trust in customers’ minds. Also, media, such as social media, can provide better services day by day with innovations and updates of their features (Khoirina and Sisprasodjo, 2018). The marketing activities of companies in the media affect their customers’ intention to buy by influencing social identity, perceived value and creating customer satisfaction (Chen and Lin, 2019). In such a situation, the company should always analyse the media environment with foresight and obtain accurate information from leading trends and changes so that they can plan to adapt media policies and strategies in the future. When new technology is introduced, the rational presupposition is to use and accept this new technology (Albarran, 2004). Therefore, the issue of AI has arisen since we were faced with a large amount of information, and it has become difficult to decide and process this information; therefore, for information processing, the need for a device similar to the human brain arose (Broussard et al., 2019; Nazari, 2020). In addition to reducing costs, this feature helps to create more cost-effective computing power (de-Lima-Santos and Ceron, 2021). AI is more than 60 years old, and in just the last decade of its development, it has changed all communications in the digital world. Due to its rapid evolution, the scientific dimensions of AI are rarely studied. Also available are AI functions: QC, searching, metadata, compliance, editing, highlights, break structure or advertising, subtitling, close captioning, supervision, presenting the news and more. Therefore, the researcher of this study seeks to explain the role of AI on the application of the cloud TAM in the media. For this purpose, the conceptual model and research hypotheses in Figure 2 are placed in six categories. Therefore, in this research, we are seeking answers to the following questions:

Q1.

What is the relationship between usability attributes and the TAM model?

Q2.

What is the role of artificial intelligence as a mediator in the model?

Contextual setting

Chemicals and industries in this area have a fundamental role in the life and progress of human society. Chemicals include a large part of the constituent materials in today’s world, and in almost every industry and process, traces of the dependence of that industry and process on chemicals can be found in its various parts. Today, the two major challenges of the chemical industry can be seen in the optimization of the industry and businesses in this field and the concerns regarding the degree of damage of industries, processes and chemical substances to the environment and the ecosystem of the planet Earth. In this regard, one of the most sensitive and worrying issues is the emission of greenhouse gases in the atmosphere, which further accelerates global warming. Fortunately, new generation technologies have been able to provide suitable solutions to optimize industries and reduce their harmful effects, along with increasing productivity and reducing costs.

In the oil industry, one of the latest technologies that has been able to greatly help improve processes, such as control and management in this industry, can be considered AI. AI is a branch of computer science that focuses on building systems and processes that are capable of performing tasks that typically require human cognitive abilities, such as learning and problem-solving. According to the needs of new industries in various areas, such as reducing production costs, energy consumption management and increasing productivity, and also the need for advanced intelligent processing in the steps and processes used in the chemical industry, the use of technologies such as AI by these industries’ experts and specialists of this industry are located. The use of solutions and tools based on AI and industry can have significant effects on factors affecting the quality of the environment, including reducing emissions and controlling the disposal of pollutants, increasing energy efficiency and optimal management of natural resources. With the development of the use of intelligent technologies in industrial and service processes, new definitions of industrial systems have been formed, including the Green Industry.

Over the past decade, the chemical industry has used AI to increase operational efficiency, reduce costs and improve customer satisfaction. According to Liao and Yao (2021), a growing number of companies are now using AI technology to reduce greenhouse gas emissions in production processes and improve energy efficiency.

Literature review

Artificial intelligence and media

In the advertising and media sectors, companies seem to be showing similar enthusiasm towards AI. A market survey in Europe found that 80% of media practitioners agreed that AI would have significant impacts on their industry (Shields, 2018). Specifically, while 62% believed AI would improve decision-making, 47% thought it would improve productivity. However, one third of the respondents also felt not quite confident in their understanding of AI and how it might be applied in their work. According to the survey, less human control (47%) and trustworthiness insights (55%) were the two biggest concerns. Overall, few people felt that AI would have a negative impact on their work or job availabilities (Shields, 2018).

Today’s audience has elevated expectations in every aspect of their digital lives, thanks largely to tech firms like Amazon that offer responsive services and active engagement. In such a data-driven and direct-to-consumer world, media need to respond faster and better to audience expectations, similar to other tech company counterparts (Raconteur, 2018). As media are now interwoven into consumers’ daily lives and technology bundles, media companies must deliver engaging individual experiences to every consumer in context, in the moment, and all the time. Consequently, human resource commitment is significant in this new reality and the solution is logically the adoption of cognitive technologies.

In addition to alleviating the volume of work, AI can make the interaction of media, content, audiences and operations faster and better. To serve audiences as individuals, media companies must understand audience sentiments/preferences towards content and characters, assess resonance and promptly align content with audience preferences. To accomplish this, media companies must acquire insights from large data sets and act on it in real time. The traditional approach of collecting and interpreting audience data is no longer agile enough for such a scenario. For instance, as the appetite for online video on multiple screens increases, AI can help media companies perform mass personalization of video content experiences more efficiently and effectively. Gentzkow (2018) concluded that the key impact of AI will be on the demand rather than the supply side of media, especially in how content is matched to consumers. The application of cognitive technologies offers better means of matching content with demand, optimizing content management and scaling the process of content delivery.

Led by tech companies, such as Facebook, Amazon, Apple, Netflix and Google, with their expansion into the media sector, AI has been applied in various areas of the media industry to make “smarter” products and services. For example, AI enables Spotify’s and Netflix’s recommendation playlists, more effective content management at media companies and even chatbots to improve customer experience. A survey found that the most common ways in which news media are using AI globally were to improve content recommendations (59%); followed by workflow automation (39%); commercial optimization, such as ad targeting and dynamic pricing (39%); and intelligent agents to help reporters find stories (35%) (eMarketer, 2018). Another survey of media’s AI adoption found that cable systems or MSO (44%) had the highest adoption rate, followed by cable networks (30%) and broadcast networks (25%) (Mayeda, 2018).

Strategically, how might AI be applied to improve a media company’s competitiveness? Because of the emerging nature and newness of AI, most literature in this area has been industry-based. Some suggested that these cognitive technologies have the potential to benefit media companies by transforming the media audience-content connection process (Berman et al., 2017; Gentzkow, 2018). Specifically, from the perspective of audience insights, AI might be used to understand social conversations and other audience sentiments and preferences (Berman et al., 2017). Natural language processing and other related AI technologies can also be applied to understand and interpret a media company’s live and archived content. For example, media content can be tagged, and rich metadata and context would enable faster, easier content management. In addition, AI can be used to integrate audience and content insights, matching audience interest and relevant content in real time to deliver personalized content and a better consumption experience. The value goes beyond developing competitive advantages, as it might help media companies identify new business opportunities as well. For example, consumer conversations about a show, storylines and characters might be integrated with other third-party content sources like events, social media and local news for other potential revenue sources (Berman et al., 2017).

Recently, more studies have begun to examine the topic of human-machine interactions or communication in the context of AI (Kietzmann et al., 2018; Mou and Xu, 2017). From the perspective of business adoption, literature has emphasized the importance of innovative employee roles and working relationships between humans and machines. From the perspective of interactions between consumers and AI-enhanced machines (e.g. chatbots and machine advice), the strategic value of such applications would largely depend on their nature and implementation. There are competing theories regarding how humans might perceive machine-mimicking human behaviour. From the uncanny valley notion that suggests the human tendency to feel uncomfortable towards a human-like machine to the media equation theory, which concluded that humans are likely to see computers, television and new communication technologies similar to real social relationships with humans (Mori et al., 2012; Reeves and Nass, 1996), the effect of AI applications for media firms, like in other sectors, remains uncertain.

The artificial intelligence functions

AI technologies vary in how they function as a communicator, from interpersonal interlocutors to content producers. Voice-based assistants, such as Amazon’s Alexa, vocally respond to human questions and requests. Embodied robots interact verbally and non-verbally with people (Peter and Kühne, 2018). Automated programmes called bots enter into text-based social media interactions by posing as human conversational partners, influencing the tone and substance of these exchanges (Ferrara et al., 2016). News-writing programmes develop narratives from raw data that appear alongside human-produced stories and cannot be easily distinguished from them (Graefe et al., 2018). What these technologies share is that by functioning as a communicator, they all step into a role that, within the conceptual confines of the communication discipline, historically has been restricted to humans. In doing so, AI more than facilitates communication: it automates communication (Reeves, 2016), as well as the social processes dependent upon it (Gehl and Bakardjieva, 2017).

AI technologies of communication are designed as something with which people exchange messages, a departure from the historical role of media developed as the means through which people communicate with one another (Gunkel, 2012). It is true that talking technologies existed before AI, such as car navigation systems; however, interaction with these devices was restricted to using a narrow range of rote commands. Technology did not adapt to the user, context or message. Interactions with AI-enabled devices and programmes are dynamic rather than static, contingent upon the messages being exchanged within a particular moment and context or upon the data being fed into the program. Some AI technologies also are responsive to individual users, “learning” about their human communication partner and adjusting interactions accordingly. Some scholars go as far as to argue that emerging technologies, such as robots, not only surpass the interactive capabilities of previous devices but may eventually push past the boundaries of human communication in their integration of multiple modalities of communication (Peter and Kühne, 2018). In design and function, AI technologies are positioned as increasingly complex and life-like communication partners.

Furthermore, research in human-computer interaction has demonstrated that when technologies directly exchange messages with people, particularly when designed with human social cues, the devices and programmes are interpreted as distinct “social actors” (Nass et al., 1994). People draw on their knowledge of human interaction to make sense of and guide their exchanges with media (Reeves and Nass, 1996). Although people know that a human programmed the machine, researchers have found that people direct their messages towards the device, not the programmer (Sundar and Nass, 2000). Research into people’s behaviour with and conceptualizations of emerging technology has produced similar results, demonstrating that people perceive robots as communicative partners distinct from humans but as social nonetheless (Edwards et al., 2016) and that people interacting with a digital assistant think of themselves as exchanging messages with a technology (Guzman, 2019). Communicative AI technologies are not only designed to function as communicators but are also interpreted by people as such.

Technology acceptance model and technology approach

In particular, TAM was built to explain and predict user acceptance of specific types of technology. Some scholars have adopted TAM in various aspects of modern technology ranging from the use of websites (Chang and Chung, 2001), web retailing (Wang et al., 2006), web browsers (Morris and Dillon, 1997), online purchase intentions (Van der Heijden et al., 2003), email (Karahanna and Straub, 1999), blog usage (Hsu and Lin, 2008), instant messaging (Turel et al., 2007), mobile technology (Hong and Tam, 2006), to ERP (Sternad and Bobek, 2013). However, there is a scarcity of studies explaining the acceptance of cloud ERP using the TAM. In this view, this study used the TAM model in examining the determinant of cloud ERP.

Much research has been done on the technology adoption model and cloud computing, each of which has examined the issue. Lubis et al. (2019) conducted a study at Toyota Astra Motor, and the researchers used the advanced TAM2 to integrate the hybrid TAM and the IS success model with IBM AMOS software. Based on the results, compatibility, perceived ease, perceived usefulness, performance, internal support and intention to use were the determining factors for Toyota's successful ERP implementation. In another study, Alhanatleh and Akkaya (2020) evaluated the factors affecting the acceptance of cloud ERP in Jordanian education. They have used structural equation modelling for analysis, and they identified the technology factor, employee factor, perceived usefulness and perceived ease of use (PEOU) as important variables to influence cloud ERP acceptance. In a review article from 2007 to 2017, Nikam and Prasad (2019) addressed the implementation of the TAM and cloud ERP model. The results of their study increased the correlation between the perception of the TAM model and ERP, which ultimately gives a better understanding of the success of cloud ERP implementation in any organization. Ahn and Ahn (2020), in their research with a comprehensive approach to the factors affecting the intention to accept cloud ERP, showed that the factors of organizational culture, regulatory environment, comparative advantage, testability and vendor locking all had a significant effect on cloud-based ERP acceptance intent, while ICT skills, complexity, visibility, data security and customization had no significant effect on the intention to adopt cloud-based ERP. In their research, Salih et al. (2021) prioritized the factors affecting the acceptance of cloud ERP and important issues related to security, usability and vendors. This research will help develop new strategies or review existing strategies and policies aimed at effectively integrating cloud ERP into the cloud computing infrastructure. It also allows cloud ERP providers to set the expectations of organizations and business owners and implement appropriate tactics. Ouaadi and El Haddad (2021), in their research on cloud accounting acceptance using TAMs, showed that some variables, such as the purpose of use, motivation, remote reporting and company size, affect the intention to use cloud accounting, while other variables have no effect, such as professional class and flexibility. Chang and Hsu (2019), in a study using the cost-benefit approach, empirically examined the intention of cloud ERP users. Based on cost-benefit analysis and the TAM, this study presents a research model to examine how benefits (perceived usefulness and PEOU) and costs (perceived problems of risk and privacy) affect organizations’ intention to change from planning to developing cloud enterprise resources. This study also calculated perceived trust and control in the field of cloud computing. Perceived usefulness, PEOU and privacy concerns significantly affect intent to change. Trust can increase perceived usefulness and PEOU and reduce perceived risk. Perceived control can also reduce perceived risks and privacy concerns. Gangwar et al. (2015), in their research, considered comparative advantage, adaptability, complexity, organizational readiness, senior management commitment and training as important variables to influence cloud computing acceptance using PEOU and perceived utility (PUt) as identified intermediate variables. Also, competitive pressure and business partner support directly affect the goals of cloud computing. In their study, Eldalabeeh et al. (2021) identified the factors affecting cloud-based accounting in the Jordanian financial sector using the TAM model. According to their SEM results, senior management support, organizational competency, service quality, system quality, perceived usefulness and PEOU were positively related to the intention to use cloud accounting. The adoption of cloud accounting had a positive effect on the use of cloud accounting.

Hypothesis development

About the characteristics of innovation and their relationship with its application and implementation, Tornatzky and Klein (1982) found that three characteristics of innovation – that is, relative advantage, compatibility and complexity – had the most important fixed relationships with the application of innovation. Every innovation is known as a superior thought to what it replaces. The degree of relative advantage is frequently expressed by economic profitability, social status and other benefits (Rogers, 2003). The complexity of any innovation includes a negative correlation with its application (Yaghubi et al., 2016). It can be stated that if adopting technology is time-consuming or requires other effort to grasp it (e.g. ordinary tasks used to be done in a briefer and more classic time), that work has become complicated. Therefore, complexity will have a negative effect on cloud adoption (Chang et al., 2007; Lian et al., 2013; Premkumar and Roberts, 1999). The technology context, according to Ahn and Ahn (2020), includes the skills related to the use of technology by employees. However, cloud computing provides a challenging environment for employees and offers a fertile ground for technological innovation. They found that technology skills do not hold a significant influence on the intention to use cloud-based ERP. Unlike Kim et al. (2017), who adopted three criteria that address the technology context, namely, economic risk, performance risk and security risk – the latter two had a significant influence on the intention to adopt cloud computing and specifically software-as-a-service, as opposed to economic risk which has no impact. Oliveira et al. (2014) examined the technological context through the organization’s readiness to host cloud computing which exerts a significant impact on cloud computing adoption. Some researchers also argue that security issues related to cloud ERP technology can have negative effects (Arinze and Anandarajan, 2010; Garverick, 2014; Weng and Hung, 2014). So, the following hypotheses are proposed:

H11A.

Technological factors will have a significant impact on PU from a cloud computing system.

H11B.

Technological factors will have a significant impact on PEOU from a cloud computing system.

One of the most important factors that play a role in the use of technologies is human factors. Factors like perceptions and attitudes of people towards information technology and their demographic characteristics are among the factors that affect people’s acceptance and use of these technologies (AbuKhousa et al., 2012; Lin and Yen, 2011; Lin et al., 2012; Yusof et al., 2008). The knowledge of decision-makers and the level of innovation is not only effective in making new decisions of decision-makers but also affects people’s ability to do things in different ways (Lian et al., 2013). The importance of the human factor has been emphasized in the studies of researchers, especially ERP systems (Lin, 2010). In addition, there are related indicators in the Davis (1989) study and the later Venkatesh and Davis (1996) study. So, the following hypotheses are proposed:

H11C.

Employee factor will have a significant impact on PU from a cloud computing system.

H11D.

Employee factor will have a significant impact on PEOU from a cloud computing system.

In general, the literature suggests that AI can enhance the industry value chain by changing relationships, reinventing business platforms and expanding the power of data. AI can also create positive impacts on workflow by improving efficiency (e.g. automated decision-making), expertise/experience/effectiveness (e.g. augmenting human experts) and innovativeness (e.g. identifying alternatives and optimization) (Kelley et al., 2018). Duan et al. (2019) stressed that using AI for decision-making, either to support/assist or replace human decision-makers, is one of the most important applications in AI history; and there are research opportunities in understanding and theorizing ways to measure AI use and impact, the role of AI in decision-making, AI system implementations and cultural/ethical/legal implications of AI applications. In addition to addressing the business value of AI, studies have also identified certain risks associated with its business adoption. For example, there are innate risks that might arise from interpretability when machines communicate with humans; verifiability issues due to statistical truth-based decisions (rather than literal or context-based decisions) made by machines; and ethical dilemmas in business automation from the perspectives of different stakeholders, such as the labour force and regulators (Brynjolfsson and McAfee, 2017; Wright and Schultz, 2018). Conceptually, scholars have suggested that the biggest advance of AI is in improving perception (e.g. image/speech recognition) and cognition (e.g. automate insurance claims) in business processes (Brynjolfsson and McAfee, 2017). Agrawal et al. (2019) further argued that AI’s biggest impact is in the prediction in an uncertain world for better decision-making. Overall, academic investigations of AI business applications have often touched on machine-human bonds/communication/collaboration, innovative business models/ecosystems, labour market disruptions and value creation in specific sectors (Bolton et al., 2018; Garbuio and Lin, 2019; Kelley et al., 2018; Valter et al., 2018).

In the advertising and media sectors, companies seem to be showing similar enthusiasm towards AI. A market survey in Europe found that 80% of media practitioners agreed that AI would have significant impacts on their industry (Shields, 2018). Today’s audience has elevated expectations in every aspect of their digital lives. In such a data-driven and direct-to-consumer world, media need to respond faster and better to audience expectations, just like other tech company counterparts (Raconteur, 2018). As media are now interwoven into consumers’ daily lives and technology bundles, media companies must deliver engaging individual experiences to every consumer in context, in the moment, and all the time. Consequently, human resource commitment is significant in this new reality, and the solution is logically the adoption of cognitive technologies. So, the following hypotheses are proposed:

H12A.

Technological factors will have a significant impact on PU through artificial intelligence from a cloud computing system.

H12B.

Technological factors will have a significant impact on PEOU through artificial intelligence from a cloud computing system.

H12C.

Employee factor will have a significant impact on PU through artificial intelligence from a cloud computing system.

H12D.

Employee factor will have a significant impact on PEOU through artificial intelligence from a cloud computing system.

Bharadwaj and Lal (2012) have admitted that the perceived usefulness (PU) for a cloud-based service can be judged on the basis of increased performance, productivity, work efficiency and service utility. Therefore, organizations which perceive that cloud computing will enable them to benefit from all these advantages will agree to adopt this solution. The firm’s operational and strategic advantages that it can reap from cloud computing are known as the system’s perceived usefulness, which relates to mobility, an efficient decrease of computing costs, ease of installation and maintenance and easy analysis of data online (Arpaci, 2017). Cloud computing is capable of delivering complete service online in such a way that users do not have to be physically present to perform the analysis and operations of data (Eldalabeeh et al., 2021). Through online connectivity, there is enhanced mobility, and because of cloud computing, firms do not need to invest significant resources to develop IS because cloud computing vendors currently install, maintain and upgrade the system, reducing the costs of IT. On the basis of the above discussion, cloud computing is expected to provide a significant advantage as highlighted by Gupta et al. (2013) and Lal and Bharadwaj (2016). So, the following hypothesis is proposed:

H21A.

Perceived usefulness (PU) will have a significant impact on attitudes towards using cloud computing.

Several research studies have pointed out the evidence on the significant effects of PEOU and the intention to use it directly or indirectly. According to Venkatesh and Davis (2000), PEOU is confirmed by factors related to intrinsic motivations and emotions in relation to system use. Bharadwaj and Lal (2012) have proposed that PEOU can be validated by interaction which is the ease and the understandability of the cloud service, less mental effort in using the solution and the ability to perform tasks, according to the organization’s guidelines. In some instances, employees work externally to an actual physical office and can still have easy data access through mobile devices, which is a great advantage (Raut et al., 2017; Zin et al., 2016). Remote access to online transactions, even in remote locations, is required by employees in the face of an increased number of online transactions, and this calls for the need for cloud computing solutions (Chiregi and Navimipour, 2018; Sabi et al., 2016). So, the following hypothesis is proposed:

H22B.

Perceived ease of use (PEOU) will have a significant impact on attitudes towards using cloud computing.

Methodology

The present study seeks to explain the role of AI as a mediating variable on the application of the TAM in the media with a cloud computing approach. The purpose of the research is applied studies and from the point of view of data collection is a descriptive survey. The statistical population of the study is the managers, deputies and experts of the National Iranian Oil Company, from which the statistical sample of 368 people, as shown in Table 2, were randomly selected.

Table 3 shows that the highest number were men (with 88%), the highest level of education to participate in this study were people with a master’s degree (with 60%), the highest work experience were people between 11 and 20 years (with 50%), and, finally, the most participants in this study to answer research questions were the company’s deputies (with 72%). The researcher also used a questionnaire (see Appendix) as a tool to collect his data based on the following existing questionnaires:

Technological factor (TF): Gangwar et al. (2015), in their very influential research, considered two factors of compatibility and complexity. This questionnaire consists of 12 questions. In their study, they stated reliability as 0.821, KMO index as 0.583 and Bartlett test as 0.000, which indicates that their questionnaire was approved.

Employee factor (EF): Lin (2010) stated that employees can increase their productivity in providing complex and technical computer solutions by using a cloud computing system. This questionnaire consists of eight questions. The convergent validity of the structures (minimum 0.5), as well as the combined reliability (above 0.70) for all indicators, were confirmed in this study.

Perceived usefulness (PU): Davis’ TAM proposed that “perceived usefulness” and “perceived ease of use” affect “attitude toward usage”; “attitude toward usage” and “perceived usefulness” affect “intention to use”; and finally, “intention to use” affect “usage behaviour”. The above-stated relationship has been validated in many research and conference papers (Alhanatleh and Akkaya, 2020). This questionnaire consists of ten questions.

Perceived ease of use (PEOU): According to TAM, potential users’ and/or users’ PEOU of an ERP system has a positive influence on their intention to use the system and attitude towards the use of the system. The relationship has been tested across different areas related to technology adoption. The above theoretical arguments have been empirically validated by various scholars (Alhanatleh and Akkaya, 2020). This questionnaire consists of 12 questions.

Attitude towards usage (ATU): Santamaría-Sánchez et al. (2010) believed that the success or failure of cloud computing has been proven in many projects. This questionnaire consists of five questions. They validated their model using four statistical models.

Artificial intelligence (AI): Since there was no standard questionnaire to measure the variables in the research, a study-based questionnaire was developed by studying the theoretical and experimental foundations and, in coordination with experts, according to the objectives of the research. This questionnaire consists of eight questions, and its validity and reliability have been confirmed.

Also, in this research, PLS 2.0 software has been used for analysis.

Analysis and results

Assessment of measurement model

SEM with partial least squares (PLS) was used for evaluation. PLS has been used to validate the causality of structural models (Costa et al., 2020). In this study, the researchers used it to analyse the research hypotheses. The PLS approach is used in small samples with a non-normal distribution that also reduces the residual variance of dependent variables (Hair et al., 2011). Also, although the research structures have been measured in previous research, validity and reliability have been evaluated separately. Table 4 presents the results of the evaluations.

To confirm construct reliability, Coelho and Henseler (2012) believe that 0.800 is appropriate. As shown in Table 5, all structural numbers are above 0.900. Cronbach’s alpha for structures is also higher than 0.70 and is therefore approved. The average variance extracted (AVE) index has been used to confirm convergent validity (Fornell and Larcker, 1981). It is clear from the results of Table 5 that all structures have a number higher than 0.600 and are approved. The result of CR > AVE and AVE > 0.5 is also valid for all structures. The quality of the structural model was then evaluated through bootstrapping and the PLS algorithm. Bootstrapping is a resampling technique that draws many subsamples retrieved from the original data set (Costa et al., 2020). For this research, 4,000 resamples were used to determine the path’s significance within the structural model. Table 5 shows the correlation of the main research structures.

All indicators load more highly on their own construct than on other constructs, without any exceptions (Hair et al., 2011). Furthermore, another criterion for evaluating discriminant validity is that the square root of the AVE should be larger than the inter-construct correlations (Chin, 1998; Fornell and Larcker, 1981). As shown by comparing the inter-construct correlations and square root of AVE (shaded leading diagonal) in Table 5, all constructs have more variance with their indicators than with other constructs. Hence, the discriminant validity of all the first-order factors and factors are supported.

The results of structural modelling are shown in Figure 3. The technological factor, according to the H11A hypothesis, has a positive and significant effect on PU (β = 0.268, ρ < 0.001). But the technological factor, according to the H12A hypothesis through AI, has a better effect on PU (β = 0.329, ρ < 0.001). The technological factor, according to the H11Bhypothesis, has a positive and significant effect on PEOU (β = 0.274, ρ < 0.001). But the technological factor, according to the H12B hypothesis through AI, has a better effect on PEOU (β = 0.319, ρ < 0.001). Human factor, according to the H11C hypothesis, has a positive and significant effect on PU (β = 0.286, ρ < 0.001). But the human factor, according to the H12C hypothesis through AI, has a better effect on PU (β = 0.343, ρ < 0.001). The human factor, according to the H11D hypothesis, has a positive and significant effect on PEOU (β = 0.223, ρ < 0.001). But the human factor, according to the H12D hypothesis through AI, has a better effect on PEOU (β = 0.347, ρ < 0.001). According to the H21A hypothesis, PU has a positive and significant effect on ATU (β = 0.453, ρ < 0.001). Also, according to the H22B hypothesis, PEOU has a positive and significant effect on ATU (β = 0.426, ρ < 0.001).

The structural model’s quality is assessed through squared multiple correlations (R²) and the model’s statistically predictive accuracy (Q2). After the validation of these two measures and assessment of the model’s quality, it was concluded that the model is valid. The latent variable’s coefficient path was analysed to study the hypotheses and if each one had statistically predictive evidence, and it was considered significant (ρ < 0.050) and the relationships between constructs were considered supported. Table 6 summarizes all the hypotheses results.

The mediator has been called an intervening or process variable. Complete mediation is the case in which independent variable no longer affects a dependent after the mediator has been controlled, making the path a direct zero. Partial mediation is the case in which the path from independent to dependent is reduced in absolute size but is still different from zero when the mediator is introduced. So, what has happened in this study suggests complete AI mediation in the model. Also, the fit indices of the research model are listed in Table 7.

Therefore, according to the results of Table 7, the model has a good fit.

Discussion

Hypotheses discussion

The current research shows that, at present, electronic learning has a very low acceptance scale at the infrastructure level. Today, the world is facing rapid changes in the field of organizational and educational systems, which require special solutions. In information technology, progress along with new technologies means increasing growth and gaining benefits from the birth of new technologies. An option that has recently been introduced in the world of information technology for the infrastructure of electronic learning systems is cloud computing technology. This technology has made a significant impact in the field of information and communication technology by changing ways of payment and costs and the use of software and hardware. The popularity of cloud computing is due to its two key features:

1. all processing needs are provided as a service; and

2. ability to provide computing and processing resources dynamically.

Before analysing the hypotheses, the research results showed that all the factor loadings are higher than 0.4, and therefore, all of them remain for analysis. Also, the validity and reliability analysis of research questionnaires indicates their approval. Also, the output results of the correlation and validity analyses of the constructs showed that all constructs have more variance with their indicators than with other constructs. Hence, the discriminant validity of all the first-order factors and factors are supported.

What was proven in this study, as in other studies, is the effect of usability attributes on the TAM model structures, which has already been obtained in researchers’ prior research (H11A-H11B-H11C-H11D-H21A-H22B) (Ahn and Ahn, 2020; Davis, 1989; Kim et al., 2017; Oliveira et al., 2014; Lin, 2010; Venkatesh and Davis, 1996). But the main topic of the article is to address the effect of AI on the model. But what was obtained about the research hypotheses (H12A-H12B-H12C-H12D) has a positive and significant effect on the mediating effect of AI in the TAM with cloud approach. de-Lima-Santos and Ceron (2021) believed that AI can make it easier to use resources without embedding unique skills. AI can also enhance new forms of participation and new products for media use (Diakopoulos, 2020; Jamil, 2020). In areas such as advertising, companies also seem eager to use AI (Chan-Olmsted, 2019). Shields’ (2018) study on the European market showed that 80% of media activists agreed with the use of AI; while 62% believed that AI influences decisions and 47% believed that it increases productivity. The results of this study also showed that few felt that AI could negatively affect performance. In addition to reducing workload, AI can speed up and improve the interaction of media, content, audience and operations (Chan-Olmsted, 2019). Gentzkow (2018) concluded that the key impact of AI is and will be on demand rather than the supply side of media, especially in how content is matched to consumers. The application of cognitive technologies offers better means of matching content with demand, optimizing content management and scaling the process of content delivery. An eMarketer survey found that the most common ways in which news media are using AI globally were to improve content recommendations (59%), followed by workflow automation (39%), commercial optimization (e.g. ad targeting and dynamic pricing) (39%) and intelligent agents to help reporters find stories (35%) (eMarketer, 2018). Another survey of media’s AI adoption found that cable systems or MSO (44%) had the highest adoption rate, followed by cable networks (30%) and broadcast networks (25%) (Mayeda, 2018).

Also, the results related to the fit of the structural model showed that the normalized chi-square index was equal to 1.54, the GFI index was equal to 0.85, the NNFI index was equal to 0.91, the CFI index was equal to 0.90 and the RMSEA index was equal to 0.057, and were confirmed. One of the most important limitations of the research was the lack of research related to media, AI and cloud computing.

Theoretical implications

AI has opened up a new space in the digital world, so that its profound impact is quite evident and has affected people’s lives. Different industries benefit from AI to varying degrees. The field of media and communications are also in the middle categories of the capacity to use AI. The main area of effectiveness of AI is in data classification and metadata extraction. It can be expressed that with the help of AI, this division can be done in real time. So that we can use this technology in the daily affairs of a television network and even in live programmes, such as news. However, the benefits of AI are not limited to producing information directly from the media. The accuracy, speed and amount of information obtained from AI tools is applied in many other branches of the media industry. Adding information from any content gives the user more searchable parameters in the content. The user’s needs are easily adapted to these parameters, and more useful and concise results are obtained. In addition, thanks to the processes performed by machine learning, it is possible to train the system and adapt it to the specific needs of the user. An example of this is by holding the image of an anonymous person; we can find it in all archived content.

Based on content analysis by AI, users can not only access all the content of a media but also specific sections of the content can be retrieved (such as special scenes in a movie, an important person talking about a particular topic or even a scene error in a sports match). Using the available editing tools, users can automatically view video, separate the existing parts and present it in the form of a new clip. The analysis of languages, translations and advances in the field of speech-to-text conversion has been able to help build an automatic translation and subtitling system, also online. AI tools are very useful in the processes of converting speech to text, marking texts and so on. By law and policy, it is sometimes necessary to delete or edit portions of existing content. AI enables us to edit or delete content based on desired parameters (such as recognizing content in images, recognizing emotions or linguistic analysis of beauty).

Conclusion

A practical view of the future

In the future, AI will have a profound impact on almost every software on the market, including enterprise resource planning systems. AI in ERP affects the controlled business nature of the system. While AI technology is still in its infancy, machine learning is making serious waves in the software market, impacting applications, such as business management, and giving them superstructure power. AI has the potential to bring about massive scale change for companies whose customer and operational data is their core business. Also, in general, every industry uses a corner of its advantages. Meanwhile, the impact of AI technology is more prominent in data-intensive and processing-intensive sectors that are most dependent on future predictions based on historical information. AI can help companies organize large amounts of data quickly and overcome limitations, then turn it into an opportunity. AI is sure to revolutionize the future of ERP solutions. New machine learning technology that appears to mimic human actions will impact any company or organization regardless of size or industry. AI can influence daily operations as part of an ERP system. This smart technology will recognize the user’s behaviour pattern and automatically perform their normal tasks. AI is reshaping ERP functions, and incorporating AI into ERP systems will make business management tools more powerful. With the advent of cloud computing, more vendors are offering cloud hosting options, making ERP solutions more accessible to SMBs. By using these systems, users no longer have to worry about the heavy costs of maintaining and upgrading systems. In addition, they can choose only the modules they need, which gives them more control over their budget. In the future, ERP users can benefit from more powerful analytics and better reporting capabilities. The future of ERP refers to a global tool that can process both structured and unstructured data. This suggests that another process that ERP managers should focus on is learning how to use big data analytics. ERP solutions will have access to all the data of different departments. Therefore, they have all the information they need to predict future trends. This makes predictive analytics a convenient feature for modern ERP platforms.

When performing any specific task in an ERP system, a specific set of procedures must be followed. AI tools can work in a similar way. For this purpose, an AI system must be programmed with a set of rules governing each action. This technology is smart enough to use these established rules to make accurate decisions and, at the same time, much faster than any human. As a result, AI enables ERP systems to automate routine responsibilities. Freeing the valuable time of employees to focus on tasks that add to business profitability is one of the other benefits of this technology. One of the main functions of AI technology is to provide intelligent outputs based on human inputs. The AI system can read users’ previous data and learn from their previous behaviour patterns. This makes ERP applications with AI provide a personalized and extensive user experience, as well as perform faster and more usefully. The latest generation of Focus ERPs has in-memory cloud computing and the ability to provide a large library of modules in finance, such as credit and collections, inventory and warehouse management, production, project costing and many more areas. Focus AIFA is a conversational API system that enables ERP to manage the entire business so that it can answer all requests at the speed of light simply by talking to it.

AI is also changing the way businesses operate. This allows industry players to use augmented reality and virtual reality features in their advertising. They can also offer an impressive list of products to dramatically improve the shoppers’ experience even before they buy. Soon, chatbots will improve conversational functionality in the retail industry. They can capture the emotions and moods of customers based on emotional responses. In short, AI enables organizations to provide uniquely personalized responses to customer requests and significantly improve customer loyalty. AI creates tremendous opportunities to enhance human capabilities in industries.

Figures

Figure 1.

Conceptual structure of individual IT acceptance models

Figure 2.

Conceptual model of research

Figure 3.

Hypotheses testing results

Table 1.

Comparison of indicators from the perspective of technology acceptance models

Model Item	TRA	TPB	IDT	TAM1	SCT	TAM2
Attitude (feeling)	*	*		*	*
Mental norm	*	*				*
Perceptual usefulness			*	*	*	*
Ease of perceptual use		*	*	*		*
Anxiety					*
Computer self-belief					*
Output quality						*
Facial			*			*
Experience						*
Volunteering						*
Visibility of results						*
Competitive pressure						*
Customer satisfaction						*
Perceptual pleasure						*
Encouragement						*
Technical support						*
Education						*

Table 2.

Statistical sample

Row	Items		No.	Percentage (%)
1	Gender	Men	324	88
		Women	44	12
2	Education	BA	63	17
		MA	219	60
		PhD	86	23
3	Experience	<10 years	54	14
		11 to 20 years	194	52
		>20 years	128	34
4	Job title	Expert	39	10
		Vice	269	72
		Manager	68	18

Table 3.

Validity and reliability of research questionnaires

Questionnaires	No. of experts	Validity	Reliability
TF	5	KMO: 0.78	0.73
		Bartlett’s test of sphericity: 0.000
EF		KMO: 0.88	0.88
		Bartlett’s test of sphericity: 0.000
PU		KMO: 0.73	0.71
		Bartlett’s test of sphericity: 0.000
PEOU		KMO: 0.79	0.70
		Bartlett’s test of sphericity: 0.000
ATU		KMO: 0.85	0.82
		Bartlett’s test of sphericity: 0.000
AI		KMO: 0.89	0.84
		Bartlett’s test of sphericity: 0.000

Table 4.

Measurement model results

Construct	Item	Loading	Internal reliability	Composite reliability	Cronbach’s alpha	AVE	Discriminant validity
TF	Tf1	0.802	0.772	0.903	0.900	0.591	Yes
	Tf2	0.771	0.719
	Tf3	0.842	0.807
	Tf4	0.726	0.699
	Tf5	0.773	0.728
	Tf6	0.812	0.743
	Tf7	0.901	0.828
	Tf8	0.800	0.783
	Tf9	0.764	0.732
	Tf10	0.719	0.667
	Tf11	0.643	0.598
	Tf12	0.697	0.601
	Tf13	0.708	0.619
EF	Ef1	0.808	0.606	0.937	0.916	0.670	Yes
	Ef2	0.846	0.657
	Ef3	0.876	0.729
	Ef4	0.789	0.637
	Ef5	0.743	0.608
	Ef6	0.839	0.760
	Ef7	0.881	0.666
	Ef8	0.759	0.633
PU	Pu1	0.776	0.674	0.962	0.893	0.667	Yes
	Pu2	0.881	0.799
	Pu3	0.909	0.843
	Pu4	0.832	0.800
	Pu5	0.826	0.764
	Pu6	0.819	0.777
	Pu7	0.765	0.739
	Pu8	0.777	0.681
	Pu9	0.770	0.701
	Pu10	0.803	0.669
PEOU	Peou1	0.917	0.883	0.961	0.917	0.768	Yes
	Peou2	0.934	0.922
	Peou3	0.902	0.829
	Peou4	0.886	0.797
	Peou5	0.808	0.736
	Peou6	0.789	0.709
	Peou7	0.793	0.711
	Peou8	0.855	0.699
	Peou9	0.888	0.782
	Peou10	0.946	0.886
	Peou11	0.834	0.743
	Peou12	0.946	0.829
ATU	Atu1	0.904	0.844	0.960	0.954	0.785	Yes
	Atu2	0.825	0.792
	Atu3	0.863	0.803
	Atu4	0.918	0.842
	Atu5	0.918	0.842
AI	Ai1	0.858	0.810	0.930	0.892	0.764	Yes
	Ai2	0.827	0.799
	Ai3	0.843	0.762
	Ai4	0.878	0.824
	Ai5	0.973	0.759
	Ai6	0.806	0.722
	Ai7	0.884	0.815
	Ai8	0.913	0.850

Table 5.

Inter-construct correlation coefficients and square root of AVE (in italic on diagonal)

Items	TF	EF	PU	PEOU	ATU	AI
TF	0.819
EF	0.552	0.862
PU	0.491	0.443	0.924
PEOU	0.451	0.369	0.563	0.964
ATU	0.623	0.346	0.490	0.556	0.911
AI	0.796	0.884	0.608	0.538	0.629	0.902

Table 6.

Hypotheses results

Hypothesize	IV		DV	Mediator		Findings	Conclusion
H11A	FT	→	PU			0.268 Positively and statistically significant	Supported
H12A	FT	→	PU	→	AI	0.329 Positively and statistically significant	Supported
H11B	ET	→	PU			0.274 Positively and statistically significant	Supported
H12B	ET	→	PU	→	AI	0.319 Positively and statistically significant	Supported
H11C	FT	→	PEOU			0.286 Positively and statistically significant	Supported
H12C	FT	→	PEOU	→	AI	0.343 Positively and statistically significant	Supported
H11D	ET	→	PEOU			0.223 Positively and statistically significant	Supported
H12D	ET	→	PEOU	→	AI	0.347 Positively and statistically significant	Supported
H21A	PU	→	ATU			0.453 Positively and statistically significant	Supported
H22B	PEOU	→	ATU			0.426 Positively and statistically significant	Supported

Table 7.

Fit indices for measurement and structural model

Type	Index	Measurement model	Recommended value
x2 test	x2	452.26
	df	292
	x2/df	1.54	≤3.00 (Bagozzi and Yi, 1988)
Absolute fit index	GFI	0.85	≥0.80 (Seyal et al., 2002)
Comparative fit index	NNFI	0.91	≥0.90 (Bentler, 1988)
	CFI	0.90	≥0.90 (Bentler, 1988)
	RMSEA	0.0573	≤0.08 (Hair et al., 2011) Note: ≤0.10 is acceptable (Rai et al., 2002)

General questions
Gender	Men
	Women
Education	BA
	MA
	Dr
History	<10 years
	11 to 20 years
	>20 years
Job side	Expert
	Vice
	Manager

Appendix

Dear expert,

The questionnaire that has been provided to you has been developed to conduct field research entitled “Explaining the effect of Artificial Intelligence on the Technology Acceptance Model in Media: Cloud Computing approach”. I request you to answer the relevant questions, according to the response range.

Responsiveness range

• I strongly disagree = 1

• I disagree = 2

• No idea = 3

• I agree = 4

• I agree very much = 5

Specialized questions

Technology factor

• Steps to complete a task in the cloud e-learning system follow a logical sequence.

• Performing an operation in the cloud e-learning system led to a predicted result.

• Screens of the cloud e-learning system were clearly organized.

• The cloud e-learning system was characterized by rapid response even at peak times.

• The cloud e-learning resources system provided relevant information for work.

• The cloud e-learning system presented the information in an appropriate format.

• The information from the cloud e-learning system was up-to-date enough for my purposes.

• The reliability of output information from the cloud e-learning resources system was high.

• The cloud e-learning system provided the information when I need in time.

• The cloud e-learning system had a modern-looking interface.

• The cloud e-learning system provided the right solution to my request.

• The cloud e-learning system gave me prompt service.

• The cloud e-learning system had a good interface to meet my needs and labour.

Employee factor

• I have experience in using handheld devices (laptops, tablets, smartphones,…).

• I have experience in using the internet.

• The cloud e-learning system is exactly what I need.

• I am sure it was the right thing to adopt the cloud e-learning system.

• Owning the cloud e-learning system has been a good experience.

• I am satisfied with the performance of the cloud e-learning system service.

• I am satisfied with the decision to work over the cloud e-learning system.

• I am pretty satisfied with the cloud e-learning resources system which has been chosen.

Perceived usefulness

• Using the cloud e-learning system improved the quality of the work I do.

• Using the cloud e-learning system gave me greater control over the activities in my work.

• The cloud e-learning system enabled me to accomplish tasks more quickly.

• The cloud e-learning system supported critical aspects.

• The cloud e-learning system increased my productivity.

• The cloud e-learning system improved my job performance.

• The cloud e-learning system allowed me to accomplish more work than would otherwise be possible.

• The cloud e-learning system enhanced my effectiveness on the job.

• The cloud e-learning system made it easier to do my job.

• Overall, the cloud e-learning system was useful in my job.

Perceived ease of use

• Overall, I found the cloud e-learning system interface easy to use.

• My interaction with the cloud e-learning system was clear and understandable.

• The cloud e-learning system required the fewest steps possible to accomplish what I want to do with it.

• Using the cloud e-learning system is effortless.

• I could use the cloud e-learning system without written instructions.

• I did not notice any inconsistencies when I use the cloud e-learning system.

• I could recover from mistakes quickly and easily over the cloud e-learning system.

• I could use the cloud e-learning system successfully every time.

• Learning to use the cloud e-learning system interface was easy for me.

• It was easy for me to become skilful at using the cloud e-learning system interface.

• I found the cloud e-learning system interface to be flexible to interact with.

• I easily remembered how to use the cloud e-learning system.

Attitude towards use

• I had a generally favourable attitude towards using the cloud e-learning system.

• I believed it was a good idea to use the cloud e-learning system for my work.

• I liked the idea of using the cloud e-learning system.

• Using the cloud e-learning system provided me with a lot of enjoyment.

• Overall, I enjoyed using the cloud e-learning system.

Artificial intelligence

• By using smartphones and their educational materials, my communication with my organization has improved.

• It has become easier to do the work of the organization by using smart devices and appliances at home.

• By using smartphones, it has become easier and easier to control the health of myself and the members of my organization.

• With the use of new communication tools, shopping, travelling, banking and going about my life have become easier.

• With the use of smartphones, the time I spend doing things has decreased.

• The use of smartphones and new communication technologies has increased the accuracy and speed of my work.

• By using relevant programmes and software in smartphones, my imagination of my surroundings has increased.

• With the use of new communication technology, the quality of the tasks assigned in my work environment has increased.