The influence of artificial intelligence on the AISs efficiency: Moderating effect of the cyber security

Abstract The study designed to identify the influence of artificial intelligence on the efficiency of Accounting Information Systems (AIS) in the Jordanian industial companies, where the descriptive analytical approach was used. The study population included all Jordanian industrial companies listed in Amman Stock Exchange, includes (55) industrial companies. The respondents includes the managers, where (275) questionnaires were distributed to the study respondents electronically, with an average of (5) questionnaires for each company, and retrieved 142 valid questionnaires for statistical analysis. The results presented that there is a significant and positive effect of artificial intelligence’s dimensions (i.e. expert systems, genetic algorithms, intelligent agents) on the efficiency of AIS in Jordanian industrial companies. But the resule revealed that neural network as an artificial intelligence’s dimensions have non-significant effect on the efficiency of AISs. The result also found that the effective application of cyber security moderates the relationship between artificial intelligence and the the efficiency of AIS positively. Given the importance of the industrial companies’ sector in the context of Jordan, the results are useful for these companies with regard to the issue of AIS efficiency and the role of artificial intelligence applications in this efficiency. The results also highlighted the importance of the effective application of cyber security in such a technological era. To the knowledge of the authors, this study is one of the first to address cybersecurity as a moderating variable on the relationship between artificial intelligence applications and AIS efficiency in industrial companies in Jordan, a developing country.


Introduction
An effective and integrated system to manage all accounting actions in light of modern computing technologies is a significant factor in businesses (Alshira'h et al., 2020;Damerji & Salimi, 2021;Lutfi, 2022a;Alsyouf, et al., 2022).Technology and innovation must be used to achieve progress and improve performance in those businesses (Alsyouf et al., 2021;Alsaad et al., 2023;Al-Okaily et al., 2020;Lutfi, 2022a;Mohammad et al., 2020).Where cases of uncertainty, the information technology development, quick advancement in science and complications in the business surroundings are among the most critical factors that have emerged from new business standards and various economic circumstances (Idris, & Mohamad, 2016), which led to the promotion of competition both locally and globally between businesses (Al-Qudah et al., 2022;Almaiah, Hajjej, et al., 2022;Lutfi, 2020;Lutfi, Farhan Alshira'h, et al., 2022b).So, the accounting profession cannot be considered in isolation from these developments (Idris, & Mohamad, 2017), at the same time if you do not take into account these developments, the accounting profession will not be able to meet the needs and aspirations of the modern business environment (Alshira'h et al., 2021a;Alshira'h et al., 2021b;Al-Qudah et al., 2022;Saad et al., 2022).Therefore, it was imperative that the accounting domain is aware of these transformations that would reorganize the economic units' work that serve it (Lutfi Al-Khasawneh, et al., 2022c).The accounting domain also needs to keep up with all these evolutions and to exploiting them by presenting the artificial intelligence model (Abdul Ridha & Al-Maamouri, 2020;Lutfi, Lutfi Al-Khasawneh, et al., 2022d).
Artificial intelligence continues to develop rapidly (Alsyouf et al., 2022;Almaiah, Hajjej, et al., 2022;Haenlein & Kaplan, 2019;Holzinger et al., 2019;Kokina & Davenport, 2017).Artificial intelligence through its dimensions participates in understanding the nature of human intelligence, by creating computer programs intended to simulate intelligent human behaviour and capable of processing operations electronically, as well as delivering internal or/and external users with the financial data and information they need in various decisions quickly and in a timely manner (Almaiah, Al-Rahmi, et al., 2022;Haenlein & Kaplan, 2019).The most significant modern ideas in communication and information technology are to develop its strategic role in building expert systems, nervous networks, genetic pamphlets, and intelligent agents (i.e., artificial intelligence dimensions) that are efficient (Rabbani et al., 2023).AISs have become a crucial part of management tools in companies and an essential resource that depends on it to promote administrative and financial operations, support decisions, and enhance the performance (Al-Hashimy et al., 2019;Lutfi, et al 2022e;Lutfi, Ashraf, et al., 2022;Norwahida & Shukeri, 2014).
Furthermore, Artificial intelligence is a very special and important topic (Almaiah, Alfaisal, et al., 2022;Almaiah, Mugahed Al-Rahmi, et al., 2022;Holzinger et al., 2019;Raisch & Krakowski, 2021).Considering developments in information technology, and the change in the performance of the accounting profession and governance, the application of artificial intelligence with its dimensions (expert systems, nervous networks, genetic algorithms, smart agents) has become required to keep pace the needs of electronic business in companies (Almaiah, Alfaisal, et al., 2022;Damerji & Salimi, 2021).The field of accounting faces significant challenges as a result of scientific advancement, technological evolution and the emergence of artificial intelligence techniques (Lutfi, Alkilani, et al., 2022;Mohammad et al., 2020;Saad et al., 2022).Consequently, Jordanian companies face difficulty in providing an accounting system that keeps pace with these new technologies because of the need to provide ready programs and to provide and train accountants who are characterized by the skill needed to deal with these programs (Lutfi & Alqudah, 2023).Besides, companies require the maintenance and updating of these programs continuously and the provision of material requirements for accountants in terms of storage devices and means, which are that it is often expensive and risky.
The remarkable development allowed opportunities for companies to improve their service level provided to customers by employing recent innovative channels, unlike what companies are used to (Alshirah et al., 2021;Al-Qudah et al., 2022;Lutfi et al., 2022g).This has made a radical shift in the way of work in all companies and has contributed to technical development in companies providing their services by conducting electronic transactions resulting in saving effort, money and time via these innovative recent channels (H.Alqudah et al., 2023;Lutfi, et al., 2022h).Meanwhile, security risks have increased, as failure to use effective cyber security procedures can lead to disruption of necessary financial services, and these are cyber-attacks that threaten the whole financial system and consequently incur immense losses when those attacks happen (Alhayani et al., 2021;Almaiah, Hajjej, et al., 2022;Razzaq et al., 2013;Suresh et al., 2022).This has led to highlighting the strategy of cyber security so that each company sets its cyber security strategy according to practices of risk management (Almaiah, Alhumaid, et al., 2022;Alrawad et al., 2022;Apruzzese et al., 2018;Dube & Mohanty, 2022).
However, the significance of the present study arises from the originality of the topic of artificial intelligence and the efficiency of AIS in light of the continuous development, and the competition for the effective application of cyber security in Jordanian industrial companies.The idea of the present study was reached due to the benefits delivered by adopting artificial intelligence in different areas, which promotes access to financial/non-financial data or information from anywhere and at any time, and helps it to readily keep pace with developments and updates in international standards.
There are several reasons to conduct this research including: 1) very few studies have addressed the artificial intelligence including its dimensions as a factor affecting accounting systems espcialy in developing countries such as Jordan.2) Most prior studies in this domain have addressed the banks sectors, while this study addressed the industrial companies' sector. 3) The main contribution of this study is investigating the effect of artificial intelligence on accounting system in light the syberscurity in the Jordanian public shareholding companies, which have neglected in prior studies.Further, the industrial sector is one of the important pillars of development in countries, as the industrial sector contributes to its effective role in the process of economic development and is characterized by having powers capable of influencing economic systems.Thus, the industrial sector assumes a great responsibility, which is to raise the standard of living of the individual in terms of absorbing them within the labor market and raising the level of their skills and experience (Ministry of Industry and Trade, 2019).Based on the foregoing, given the importance of the Jordanian industrial sector and its role in strengthening the Jordanian economy, Thus, this study seeks to know the impact of artificial intelligence on the efficiency of accounting systems in Jordanian public shareholding companies in light of the effective application of cyber security.This article is designed as follows: section 1 reviews the introduction, section 2 presents the literature review and develops the hypotheses, section 3 will a/ddress the methodology used, and section 4 will show the main analysis findings.Lastly, section 5 will present the discussion and conclusion.

The effect of artificial intelligence on the efficiency of accounting systems
A few years ago, artificial intelligence was out of reach, as we only saw robots and modern technology devices in fictional movies (Damerji & Salimi, 2021;Raisch & Krakowski, 2021).But today, with technological development artificial intelligence applications have become indispensable in our daily lives (Almaiah, Alfaisal, et al., 2022;Holzinger et al., 2019;Kokina & Davenport, 2017).Modern technology machines that perform most daily tasks are becoming more efficient and effective over time using modern artificial intelligence techniques (Almaiah, Alfaisal, et al., 2022;Askary et al., 2018).Also, modern accounting applications have developed thanks to artificial intelligence, which has a positive impact on accountants, as it saves them time and performs calculations in a short time (Alkan, 2022;Almaiah, Al-Otaibi, et al., 2022;Faccia et al., 2019;Mohammad et al., 2020;Zhang et al., 2023).As the collection and analysis of data and financial accounts are one of the main tasks of the accountant, in the past the accountant used to perform these tasks manually and without any means, but today there are many applications of artificial intelligence that have a positive impact and are used in AISs (Hashem & Alqatamin, 2021;Ionescu, 2019).
Furthermore, the goal of artificial intelligence in the accounting field is to increase the efficiency of computational operations in ways that help to make the best decisions (Almaiah, Ayouni, et al., 2022;H. M. Alqudah et al., 2019a;Damerji & Salimi, 2021;Khassawneh, 2014), as artificial intelligence has the ability to extract information and data with high accuracy at the click of a button, which if done by the accountant may take him several days, not to mention the human error contained it happens (Faccia et al., 2019).Artificial intelligence organizes and analyzes accounting information, which makes it easier for accountants, because of its ability to provide the most accurate financial information (Damerji & Salimi, 2021;Raisch & Krakowski, 2021;Zaitoun & Alqudah, 2020).Through this information, the accountant can build a strategy and develop a comprehensive plan for all financial assets.Artificial intelligence systems have an effective impact on AISs, as they are constantly evolving to keep pace with the continuous changes in order to be able to provide information with high efficiency as an alternative to human efforts (Alkan, 2022;Askary et al., 2018).
Several researchers e.g., Kokina and Davenport (2017), Faccia et al. (2019), Zhang et al. (2023) confirmed that artificial intelligence applications are important and necessary in many fields.As those applications represent an urgent necessity for economic units that cannot be dispensed with, in achieving many advantages, the most important of which are: calculating the decisionmaking process, solving all administrative problems, reducing costs, improving quality and other advantages, which has a major and direct role in enhancing the competitiveness of enterprises and ensuring their survival and continuity.Al-Shatnawi et al. (2019) found that the usage of artificial intelligence applications has a positive influence on improving the quality of accounting information in Jordanian public shareholding companies.And in the study by Bouzerb and Sahnoun (2019), which was entitled: "Artificial Intelligence and its applications in the banking sector" concluded that the application of artificial intelligence in the banks sector has many benefits, e.g., reducing costs, increasing employees' satisfaction, and acquisition their loyalty (N.F. AlQudah et al., 2022).Also, indicating that the applications of artificial intelligence help in reducing the negative aspects related to banking operation, such as cases of money laundering, fraud, and the disappearance of human errors.In another study, Al-Dalahma et al. (2019) entitled "The Impact of Artificial Intelligence Applications on the Accounting Profession," the results confirmed the significant influence of using artificial intelligence technology on the accounting domain.While Boua'a (2019) found that artificial intelligence applications is an inevitable strategic technology that assest in obtaining greater productivity and innovative opportunities to reach competitive advantage for many organizations.With artificial intelligence, organizations can accomplish more tasks in less time by supporting its applications.
Whereas, artificial intelligence includes the following types: Expert system -which are distinct information program that purpose to simulate the logic of human scientists, qualified people and experts.The information programs' value that are the guarantors of the effectiveness of the expert system are one of the concerns of the computer specialist.Expert systems must have several characteristics that must be available in order to be able to perform their work to the fullest.One of these characteristics is that expert systems are able to perform tasks as humans do (Qasaimeh et al., 2022).And that expert systems provide and preserve rare human knowledge and expertise, and finally, the use of expert systems should be easy in all fields with the ability to identify appropriate solutions to the problems under study and draw appropriate results for decision-making (Wan & Mahamud & Norita, 2010).
Neural network: it is known as one of the most ancient artificial intelligence techniques adopted in computer software.These networks are able to operate like the human brain.Naritatsu et al. (2013) believes that neural networks work like the nerves that designed in the human body.Neural networks can be described as a thinking model based on the human brain, as the brain consists of a large group of neurons that operate basic information (Graupe, 2007).Where the uses of neural networks in accounting systems can be summarized through the following points: (Remo, 2019).1) Neural networks are used in the areas of forecasting, especially in the preparation of estimated budgets, as they begin with the process of forecasting the size of sales, and they can also be used in the forecasting process in preparing estimates for the state's general budget from forecasting the value of expenditures for the next year.2) The use of neural networks in estimation processes, the most important of which is the estimation of provisions, such as the allowance for doubtful debts.3) Using neural networks to rationalize many decisions, the most important of which is the rationalization of the selection process between alternatives between investment projects.4) Using neural networks to detect cases of fraud and fraud.Thus, artificial neural networks are the most important new techniques and models in risk management in industrial companies, especially the risks of liquidity and financing (Lutfi, Alrawad, et al., 2023).
Genetic algorithm: is considered a computer program that simulates biological processes to analyze the evolutionary systems' problems.Genetic algorithms have become one of the most effective ways to deal with complex investigation and research issues, and they are described as genetic because they depend on simulating the work of genetic genes in order to reach the best and optimal solution (Qasaimeh et al., 2022).Where (O'Brien, 2011) defines genetic algorithms as methods that assist in producing solutions to specific issues, using methods suitable to their environment, and they are used in the areas of financial operations and control of the resources.

Intelligent agents:
A programmed activity that performs a set of operations on behalf of another user or program with a certain degree of independence.Intelligent client technology is one of the main solution technologies to deal with the problem of information overload caused by the development of a fully networked business environment (Hashem & Alqatamin, 2021).The smart agent contains of many of the following elements that interact between them, and these elements can be perception: that is, the knowledge of the data that the agent receives through sensors.Reaction: Any events issued by the agent.logical agent: It is the agent that behaves properly, that is, every row of the function table contains correct data (Latifa, 2017).The smart agent can be tasked with reading e-mail or sorting it as sales agents' reports, and as an example, searching for the best sales deal executed during the last period by the company's branches (Yassin, 2012).
In the context of the current study the AISs in the Jordanian industrial companies as any filed affected by the applications of the artificial intelligence.There are many characteristics and advantages of artificial intelligence that may support AISs in industrial companies.Qasaimeh et al. (2022) mentioned that one of the characteristics of artificial intelligence is that it helps in solving problems existing due to the absence of complete information.It helps to think, comprehend, and the capability to obtain and use knowledge.Also, the ability to learn from previous experiences and employ them in new situations.Artificial intelligence has the ability to respond quickly to new situations and conditions and to deal with difficult and complex cases (Hashem & Alqatamin, 2021).Among the characteristics of artificial intelligence is the ability to distinguish the relative importance of the elements of the presented cases, and the ability to visualize, create, understand and perceive visual matters.Finally, the ability to provide information to support administrative decisions.Therefore, AISs are greatly affected by the use of artificial intelligence applications.This leads to the following hypothesis: The AISs' efficiency in the Jordanian industrial companies affected positively by applaing artificial intelligence.
From this main hypothesis, the following sub-hypotheses branch out: H 1.1 : The AISs' efficiency in the Jordanian industrial companies affected positively by applaing expert systems.
H 1.2 : The AISs' efficiency in the Jordanian industrial companies affected positively by applaing neural networks.
H 1.3 : The AISs' efficiency in the Jordanian industrial companies affected positively by applaing genetic algorithms.
H 1.4 : The AISs' efficiency in the Jordanian industrial companies affected positively by applaing intelligent agents.

The moderating effect of cyber security
In general, companies are constantly working to protect their business, interests, and economic activities, local and global, in line with the growth and continuous development of information technologies (Alsyouf & Ku Ishak, 2018;Alsyouf et al., 2021;Alsyouf et al., 2023;Lutfi, Saleh, et al., 2022i).In the context of this study, Jordanian industrial companies that use information and communication technology and that deal with financial information and personal data are more vulnerable to cyber-attacks (Apruzzese et al., 2018;Razzaq et al., 2013).Many companies in the world are exposed annually to these attacks, and there are many companies that have been subjected to cyber-attacks and trying to access AISs and stealing data and money, and this is what exposed them to losses estimated at millions (Alhayani et al., 2021;Dube & Mohanty, 2022).Therefore, the importance of cyber security for companies and the factors for the success of cyber security in companies must be emphasised.
Cyber security is defined as protecting networks, information technology systems, operational technology systems and their components, hardware and software, and the services they provide and the data they contain from any penetration, disruption, modification, entry, use or illegal exploitation (National Authority for Cyber security, 2018).
The association between technology and security has become interdependent, with the opportunity of exposing strategic interests of an electronic nature to cyber threats (Dube & Mohanty, 2022).The continuous growth in the use of information systems' applications, and the emergence of E-commerce and E-government, all require a secure information environment (H.M. AlQudah, 2015).Thus, cyber security has become a global issue that needs flexible strategies that adapt to continuous changes, whether in the mechanisms or tactics of security in facing the continuous development of risks (Alhayani et al., 2021).The elements of information security are in themselves goals of cyber security, and they are elements that must be available to protect the information, and these elements are Confidentiality: which means preventing access to information except by authorized persons only, whether when storing or processing it or when transferring it through means of communication, as well as determining the validity of modification deletion and addition.Integration and content integrity: It is intended that the information be correct when it is entered, as well as during its transfer between devices in the network, and to ensure that it has not been affected by any change, by using a set of methods and systems.Availability and Availability: It means that the information remains available to the user and can be accessed at any time and that this is not disrupted as a result of a defect in the systems for managing information and databases or means of communication (Mahmaod et al., 2023).
Based on aforementioned, this study proposed the effective application of cyber security as a moderator factor btween the AISs' efficiency and the artificial intelligence.The continuous and accelerating technological developments that took place in the economic fields, the large size of companies and the expansion of their activities, led to the production of huge amounts of various data and information.Therefore, there was a need to use AISs and take advantage of the enormous and multiple capabilities that characterize these accounting systems.At the same time, companies have taken advantage of artificial intelligence to support and enhance the information systems and services they provide to their customers.Hence the need for an effective application of cyber security system to protect the financial operations and commercial activities of those companies has appeared (Alhayani et al., 2021).The important functions of cyber security are: to identify internal and external threats to cyber risks and protect sensitive information in the company from hacking operations; identify the entities with the ability to use the information and access the information and communications technology environment; and finally, take the necessary corrective measures to control and reduce the negative effects of electronic risks (Central Bank of Jordan, 2018).Therefore, this study argues that the effective application of cyber security could moderate the effect of artificial intelligence on AISs' efficiency.That is, applying artificial intelligence (e.i., expert systems, neural networks, genetic algorithms, intelligent agents) will increase the efficiency of AISs among Jordanian industrial companies, while different levels of effective application of cyber security may affect the efficiency of AISs positively or negatively.This leads to the following hypothesis:

Expert
The effective application of cyber security moderates the effect of artificial intelligence on AISs' efficiency in Jordanian industrial companies

Research model
As shown in Figure 1, the research model of this study encompasses three latent variables, namely: artificial intelligence (e.i., expert systems, neural networks, genetic algorithms, intelligent agents) (independent variable), AISs' efficiency (dependent variable), and effective application of cyber security (moderator variable).

Research methodolgy
The current study aims to identify the impact of artificial intelligence on the efficiency of AISs in Jordanian public shareholding industrial companies and to know the impact of the effective application of cyber security on the relationship between these two variables as a moderator variable.The study population consists of managers of administrative departments in Jordanian industrial companies listed in Amman Stock Exchange, which are (55) companies.Due to the small population, the current study took the entire population as a study sample.The questionnaires were distributed to 275 financial, marketing, financing, Human resource and information technology managers in the Jordanian public shareholding industrial companies with an average of (5) questionnaires for each company.142 usable responses were received from respondents.The reason why the Jordanian public shareholding industrial companies were selected as a population group for the current study is that these companies have a higher probability of having AISs, while most other industrial companies such as small and medium-sized companies do not have AIS systems, because of its small size and capital.
The G-Power tool was utilized to calculate the minimum sample size needed for this study.The parameters used in G-Power were as follows: an error type (a) of 0.05, an effect size of 0.15, a power of 0.95, and 5 predictors.It was determined that a minimum sample size of 59 participants was required.In total, 142 valid questionnaires were collected and used for further analysis.The questionnaires distributed amounted to 275, and out of those, 142 questionnaires were received and deemed valid, resulting in a response rate of 51.6%.This response rate meets the statistically acceptable criteria (A.Alshira'h et al., 2023;Lutfi, Nafeth Alkelani, et al., 2022j;Lutfi et al., 2020;Sekaran & Bougie, 2016).
The questionnaire was used as a main tool for collecting information about the study sample and obtaining data that express their perspective about the variables in the study context, to comprehend and explain the association between respective variables.

Measures
The questionnaire items were derived from previously validated and tested surveys.These measures were employed to assess the dependent, independent, and moderator variables within the questionnaire.The dependent variable represents the level of effectiveness of AIS (Artificial Intelligence Systems) in Jordanian industrial companies.The independent variables indicate the degree of utilization of various artificial intelligence applications (such as expert systems, neural networks, genetic algorithms, and intelligent agents) in Jordanian industrial companies.The moderator variables reflect the extent of implementation of cyber security measures in Jordanian industrial companies.Table 1 displays the measurements of each variable examined in the study.

Consructs Measurments
Expert systems Six items measured this construct adapted from Qasaimeh et al. (2022), Including Industrial companies work on periodic maintenance of the devices used by them.
Industrial companies have sufficient experience to use all different computer programs.
Industrial companies have devices and equipment that are suitable for the operation of expert systems.
Administrative and financial decisions are taken based on the reports of expert systems in industrial companies The Industrial companies' use of expert systems contributes to reducing potential risks from cyber attacks The Industrial company's use of expert systems contributes to increasing the efficiency of accounting systems

Neural networks
Seven items measured this construct adapted from Qasaimeh et al. (2022), Including Industrial companies work on periodic maintenance of the devices used by them.
Industrial companies have sufficient experience to use all different computer programs.
Neural networks support industrial companies' work.
Industrial companies have a qualified and trained human cadre to deal with neural network technology.
Industrial companies provide the necessary financial allocations to cover the expenses of applying neural network technology.
Industrial companies have advanced systems in operating neural network technology.
Industrial companies rely on the techniques and applications of neural networks in predicting cyber risks.
Industrial companies have advanced computer devices and equipment that are compatible with neural network technology.
Neural networks provide industrial companies with multiple options due to their great ability to analyze information and its risks.

Genetic algorithms
Four items measured this construct adapted from Qasaimeh et al. (2022), Including Genetic algorithms help industrial companies find quick solutions to rapid and sequential developments.
Genetic algorithms facilitate industrial companies' work in reaching quick results when there are various and complex cyber risks.

Genetic algorithms help the industrial companies perform complex calculations to get quick results
Genetic algorithms assist industrial companies in implementing effective security controls aimed at reducing cyber risks.
intelligent agent Four items measured this construct adapted from Qasaimeh et al. (2022), Including The intelligent agent helps industrial companies to make the right decision based on their stored knowledge base.
The intelligent agent contributes to reducing the amount of time it takes for industrial companies to achieve cyber governance.
The intelligent agent acts on behalf of the bank in making decisions in certain predetermined cases.

The use of intelligent agents increases the efficiency of accounting systems
The intelligent agent contributes to providing industrial companies with accounting data and information that serve the bank's objectives. (Continued)

Consructs Measurments
Cyber security Nine items measured this construct adapted from Al-Afeef et al.
(2020), Including Industrial companies' management is constantly updating its cybersecurity policy.
Industrial companies' management has a comprehensive record of cyber risks Industrial companies' management conducts a continuous monitoring process for the level of cyber risks Industrial companies' management conducts a comprehensive survey of the internal and external environment to identify cyber risks Industrial companies' management takes the necessary corrective measures to control cyber risks and limit their negative effects.
Industrial companies' management relies on effective communication channels between the concerned parties, which include responding optimally to cyber-attacks and limiting their negative repercussions.
Industrial companies' management holds training programs and workshops aimed at educating human cadres on the provisions of the cybersecurity policy.
Industrial companies' management evaluates the adequacy of their cyber security policy.
Industrial companies' management adopts procedures aimed at enhancing their ability to manage the risks of their information technology environment.

AISs Efficiency
Nine items measured this construct adapted from Al-Jaber (2020), Including the AISs that are applied in industrial companies enable raising the efficiency and ability of employees to develop and advance.
The AISs applied in industrial companies contribute to providing information that is consistent with the customer's needs.
The AISs applied in industrial companies contribute to providing information that contributes to reducing risks.
The AISs applied in industrial companies can provide the necessary information to make the appropriate decision related to the desires and needs of the customer.
It enables the AISs that are applied in industrial companies to adapt to the internal and external environment in order to raise the efficiency of their performance.
The AISs that are applied in industrial companies employ modern hardware and software to raise the efficiency of their performance.
The AISs that are applied in industrial companies provide information about training programs to raise the efficiency of employees in them.
The accounting systems applied in industrial companies encourage the creation of the appropriate environment to create a spirit of creativity and innovation among its employees.
The accounting systems applied in industrial companies achieve the lowest cost compared to using other alternatives.
The AISs applied in industrial companies achieve flexibility in the changing environment.

Common method bias
While self-reported data may include exaggerated or incomplete information, it is widely defended in the literature because it is often the sole available method for many research categories (Lutfi, et al., 2023a).To address the potential bias caused by this method, several measures were taken.For example, the data was collected in two stages, ensuring respondent anonymity.Additionally, the questionnaire items were randomized to prevent easy identification of antecedents and outcome variables.To confirm the absence of common method bias, an exploratory factor analysis with an unrotated solution was conducted.The results of this statistical test, known as the Harman single-factor test, indicated that only 37.84% of the variance was explained by a single factor, providing further evidence of the lack of common method bias.These steps align with previous studies (H.Alqudah et al., 2023;Lutfi, Alkelani, et al., 2022;Lutfi et al., 2022k).

Reliability and validity
Only a well-designed, well-organized, and dependable research instrument assures the validity of the study findings.Many dependability tests are introduced by specialists for this reason.The researcher performed a reliability analysis known as the "Cronbach Alpha Test" for this purpose, which authenticates the research instrument if the Cronbach Alpha value is larger than 0.07 (Sekaran & Bougie, 2016).This test is introduced primarily to assess the dependability of the research tool, which further assures the authenticity of the research instrument and results.As a result, the reliability analysis of the research instrument used in this study found that it is adequately reliable, that is, found that the result of Cronbach Alpha values for all studied variables is more than 0.07.
Further, the validity test aims to verify the integrity of the linguistic formulation of the study tool, the clarity of its meanings, the interdependence of its paragraphs, its objectivity, and its affiliation with the dimensions it expresses, in a way that guarantees the achievement of the objectives of the study.The arbitrators are those with experience and competence who are most capable of issuing a judgment on the validity of the content of the study tool, and for this reason, the questionnaire was presented to a group of faculty members specialized in the subject of the study in some Jordanian and Arab universities.

Data analysis and results
The "Statistical Package for Social Sciences (SPSS 25)" were used for the data analysis of this study.Data cleaning, descriptive statistics, multi-collinearity, multiple linear regression and hierarchical regression tests were conducted by using SPSS.
For demographic profile of respondents, four questions were directed to demographic data as age, education level, years of experience and job title.
As shown in Table 2, the prevalence of the respondents were within the age group of approximately 41-60 years (91%), and the vast majority of the participants had a bachelor's degree (73%) while 25% of them had a master's degree.The majority of the respondents had at least 8 years of working experience (92%).Consequently, the demographic data of the respondents revealed that they had acceptable knowledge and experience to participate in the survey and deliver trustworthy data for this study.
For descriptive statistics, as shown in Table 3 the mean scores of the variables were upper the mid-point on the one to five-scales.We classify the five-point scale into three classifications: low, medium and high scales.Scores smaller than 2.33 are considered low; scores higher than 3.67 are considered high; while scores between 2.33 and 3.67 are considered moderate (Hair et al., 2016;Sekaran & Bougie, 2016).As shown in Table 3 the mean for the study variables ranged between 3.46 and 4.11.That means all study variables have a very good mean level.More precisely, the result demonstrates that from the perspective of the managers in Jordanian industrial companies, there is a high level of AISs ' efficiency, and also a high level of artificial intelligence applications, and there is a Moderate level of effective application of cyber security in those companies.
To assess the presence of multi-collinearity among the study variables, both the variance inflation factor (VIF) test and the tolerance test were utilized.These tests aimed to confirm that there were no issues of multi-collinearity.The results of these tests can be found in Table 4. Multicollinearity refers to a situation where there is a strong linear correlation between two or more independent variables in a multiple regression analysis (Alrawad et al., 2023;Hair et al., 2014;Lutfi, et al., 2023b).The presence of multicollinearity can lead to distorted tests of statistical significance and inaccurate estimates of regression coefficients (Pallant, 2020).
According to Hair et al. (2016), "multicollinearity is not an issue when the value of VIF is less than five and tolerance is above 0.2".The results offered in Table 3 shows that there is no multicollinearity among the study variables.That is, all tolerance values are between (0.551 and 0.817) which are more than 0.20, and VIF values were between (1.348 to 1.798), which are less than 5. Thus, no multi-collinearity issue among the endogenous variables.
For Multiple Linear regression test, Table 4 depicted the multiple regression coefficients of the artificial intelligence (i.e., expert systems, neural networks, genetic algorithms, intelligent agents general) as an independent variable to the AISs' efficiency as a dependent variable.
All variables are significant where (p < 0.01).In terms of t-values, the highest t-value was discovered for the independence, where (t-value = 8.453).This indicated that independence of internal auditor's variable made the strongest contribution to explain the dependent variable (the quality of electronic internal audit).The lowest t-value indicated that internal auditors with electronic qualification affect the the quality of electronic internal audit (t-value = 3.452).The results of the testing hypotheses indicated that the independent variable significantly affect the dependent variables (Hair et al., 2016).Hence, all hypotheses are accepted (see Table 4).
It is obvious from Table 5 that there is a significant correlation between artificial intelligence and the efficiency of AISs, where the value of the determination coefficient was (R 2 = 0.419), and this indicates that artificial intelligence explained (41.9%) of the change in the efficiency of AISs and that its value (58.1%) is attributed to other factors.
Table 5 shows the values of the regression coefficients for the dimensions of artificial intelligence, where the T-value of Expert systems' dimension was (3.574) and at the level of significance (Sig = 0.00), which is less than 0.05, which indicates a positive significant effect of expert systems on the efficiency of AISs.The T-value for the dimension of neural networks was (1.072) and at the level of significance (Sig = 0.343), which is more than 0.05, which indicates that there is no significant effect of neural networks on the efficiency of AISs.The T-value of the genetic algorithms dimension was (2.778) and at the significance level (Sig = 0.004), which is less than 0.05, which indicates that there is a significant effect of genetic algorithms on the efficiency of AISs.The T-value of the intelligent agents' dimension was (3.214) and at the significance level (Sig = 0.00), which is less than 0.05, which indicates a positive significant effect of the intelligent agents on the efficiency of AISs.
For the Hierarchical Regression test, Table 6 shows the hierarchical regression result of the second Hypothesis which proposed that: the effective application of cyber security moderates the effect of artificial intelligence (i.e., expert systems, neural networks, genetic algorithms, intelligent agents) on the AISs' efficiency in Jordanian industrial companies.
It is clear from the results of Table 6, that the hierarchical regression is based on two models, where the results of the first model represented the effect of artificial intelligence on the AISs' efficiency in Jordanian industrial companies, as it was found that there was a significant effect of the combined dimensions of artificial intelligence on the efficiency of the AISs, as the value of (ΔF = 39.156) at the level of significance (Sig ΔF = 0.000), which is less than 0.05, and the value of the coefficient of determination R 2 , which amounted to (0.419), indicated that (41.9%) of the change in the efficiency of the AISs can be justified through the dimensions of intelligence combined artificial.
In the second model, the moderator variable (cyber security) was added to the regression model, as the value of the determination coefficient R 2 increased by (10.7%), which is statistically significant, as the value of ΔF reached (47.679) at a significant level (Sig ΔF = 0.000), which is less than From 0.05, the value of the coefficient of B at the dimension (cyber security) was (0.364), and at a significant level (Sig.= 0.000), and this indicates the difference in the significant impact of the dimensions of artificial intelligence on the efficiency of the AISs due to the difference in the cyber security.Hence, the main second hypothesis is accepted.

Discussion and implications
Some special practical and theoretical implications are shown in the present study.Theoretically, there are very few studies that have addressed the applications of artificial intelligence in accounting systems.This study strongly supports the adoption of artificial intelligence in promoting AIS efficiency among industrial companies.Furthermore, in prior studies on AIS, scholars have focused on the factors affecting AIS efficiency among companies, rather than examining factors that could facilitate the implementation of such systems (such as artificial intelligence), especially during and post the COVID-19 pandemic.Additionally, studying cybersecurity as a moderating factor between respective factors in the Jordanian industrial companies' context is a fascinating study that can provide insights into diverse countries, especially developing ones.Moreover, by reviewing prior studies in the context of AIS efficiency, it was found that the current study is unique in dealing with artificial intelligence and cybersecurity as moderating factors affecting AIS efficiency.Furthermore, very few studies have addressed these factors in the context of Jordanian industrial companies.Our study aims to fill this gap by investigating the determinants of "AIS efficiency" that might illustrate the efficiency of this technology in the context of Jordanian industrial companies after the COVID-19 pandemic.
The present study provides support for the effect of most dimensions of artificial intelligence on AIS efficiency in Jordanian industrial companies.Three out of the four dimensions of artificial intelligence, namely expert systems, genetic algorithms, and intelligent agents, were found to have a significant positive influence on AIS efficiency in Jordanian industrial companies.However, one dimension of artificial intelligence, namely neural networks, was found to have an insignificant influence on AIS efficiency.First, we discuss the results regarding the dimensions of artificial intelligence as independent variables.Expert Systems play a pivotal role in the efficiency of AIS within Jordanian industrial companies by providing valuable decision support to managers and employees involved in the efficiency process.They can analyze complex data, evaluate different scenarios, and provide recommendations based on predefined rules and knowledge.This helps stakeholders make informed decisions regarding the implementation and integration of AIS within the organization (H1.1).Expert Systems also assist in identifying potential risks and challenges associated with AIS efficiency within Jordanian industrial companies.By analyzing historical data, expert systems can identify patterns, trends, and potential issues that might arise during the implementation phase.This allows organizations to proactively address these risks and develop mitigation strategies, ensuring a smoother adoption process.The findings suggest that if the beliefs of industrial companies that have expert systems are consistent and compatible with AIS, a positive mindset or perception of AIS efficiency is likely to occur, ultimately resulting in the efficiency of AIS within the companies.This finding aligns with prior literature (Qasaimeh et al., 2022).
For genetic algorithms they play significant role in the efficiency of AIS in the Jordanian industrial companies (H1.2).Neural networks are used in the areas of forecasting, especially in the preparation of estimated budgets, and they can also be used in the forecasting process in efficiency of AIS.The findings suggest that if the neural networks are available among the industrial companies, a positive perception of AIS is likely to occur, ultimately resulting in the efficiency of AIS within the companies.This finding aligns with prior literature (Remo, 2019).Further, intelligent agents are an important part in the efficiency of AIS in the Jordanian industrial companies (H1.4).Intelligent client technology is one of the main solution technologies to deal with the problems caused by the development of a networked business environment, and intelligent clients can continuously learn and adapt based on user feedback and evolving business requirements, and suggest enhancements to the AIS efficiency process.By leveraging the capabilities of intelligent clients, Jordanian industrial companies can continually refine their AIS implementation and achieve better outcomes.The findings suggest that if the Intelligent clients are available among the industrial companies, a positive perception of AIS is likely to occur.This finding aligns with prior literature (Hashem & Alqatamin, 2021).
Second, we discussed the moderator effect of cybersecurity on the relationship between artificial intelligence and the efficiency of AIS in the industrial companies (H2).The cyber security is a global issue that needs flexible strategies that adapt to continuous changes, whether in the mechanisms or tactics of security in facing the continuous development of risks.Hence, the effective application of cyber security can moderate the effect of artificial intelligence on AISs' efficiency.That is, applying artificial intelligence (e.i., expert systems, neural networks, genetic algorithms, intelligent agents) will increase the efficiency of AISs among Jordanian industrial companies.The result confiremed that, the high level of cyber security will increase the effect of artificial intelligence on the AISs' efficiency.
Other results are not in line with the prior studies.We did not find neural networks to have a significant influence on AIS efficiency in Jordanian industrial companies.In this vein, previous research indicates that neural networks positively affect the efficiency of technology (Remo, 2019;Graupe, 2007).Nevertheless, we found that there is no relation between neural networks and AIS efficiency.One justification could be because of the time and the context of the research.Furthermore, this unexpected result could be because of the context of our study, which was the industrial companies in regard to AIS efficiency and could be different from another context.That is to say, concerning the neural networks as one of the artificial intelligence dimensions, the majority of industrial companies' managers (as respondents) did not see neural networks as a key determinant of AIS efficiency in the Jordanian industrial companies but did see other dimensions as a significant factor.This paper proposed a moderator variable not examined in previous studies i.e. cyber security.Cyber security moderated positively the influence of artificial intelligence on AIS efficiency.The high level of cyber security provides safe and secure areas that help institutions to use AIS efficiently.
This study contributes to the literature on artificial intelligence, AIS, and cybersecurity by presenting evidence that artificial intelligence dimensions (i.e., expert systems, genetic algorithms, and intelligent agents) have a significant effect on AIS efficiency in the context of industrial companies in developing countries.Additionally, this study offers a unique contribution by introducing cybersecurity as a moderator factor in the industrial companies' context.Moreover, the present study focuses on the industrial sector post the challenges that posed by the COVID-19 pandemic, while previous studies have predominantly addressed the service or financial sectors, and that too primarily in developed societies.
Practically, empirical results imply that Jordanian companies should pay close attention to their AISs, artificial intelligence and its applications, and be aware that they must allocate the needed resources to operate the AISs effectively.Having effective AISs helps to reduce efforts, costs and time to conduct financial activities in Jordanian companies.In addition, Jordanian companies give great attention to the effective application of cyber security to protect their activities.These findings may help decision-makers in Jordan set the legislation that could assist in adopting artificial intelligence and AIS applications among Jordanian companies.

Conclusion
In the present study, the moderating effect of cyber security on the relationship between artificial intelligence and the AISs' efficiency in Jordanian industrial companies was analyzed.The current study extended the literature related to AISs by employing artificial intelligence, as an antecedent factor to the AISs' efficiency, and also by using cyber security as a moderator variable among these factors.The present study's results support theoretically and empirically the relationship between artificial intelligence and AISs' efficiency in Jordanian industrial companies, and the role of cyber security as a moderator variable in such a suggested model.Figure 2 summarises the SPSS analysis findings of the study model.
The results of the current research indicate a significant positive influence of artificial intelligence on AISs' efficiency (with a value of p = 0.000), which is agreeable with the earlier literature in the field despite its scarcity (e.g., Hashem & Alqatamin, 2021;Qasaimeh et al., 2022).In this vein, it is recommended that companies should convince adopters of artificial intelligence to assist in enhancing the efficiency of AISs.In an effort to increase the adoption of artificial intelligence among companies, also recommended overcoming concerns about cyber attacks through the effective application of cyber security.This, in turn, may influence users' decisions about artificial intelligence adoption.The results also show that the effect of artificial intelligence on AISs' efficiency in Jordanian industrial companies is moderated positively by the effective application of cyber security.More precisely, the result confirmed that with a high level of effective application of cyber security will increase the effect of artificial intelligence on AISs' efficiency (withe value of P = 0.000).

Limitations and future research
There is no research free from limitations.The present study has some limitations.First, this study focused on the effect of artificial intelligence on AISs' efficiency.Future research can use other factors and compare the different aspects to make additional contributions.Second, this study relies on data from Jordan.Future studies recommend drawing from other countries to understand the effects of cultural differences on the research context.Third, the applications of artificial intelligence are still new in Jordan, especially in the field of accounting systems.Further research could examine the factors affecting adopting those applications.Fourth, the present study has addressed Jordanian listed industrial companies as a study population.Future research could be conducted in all industrial companies to be able generalize the results.Fifth, the study establishes associations between artificial intelligence dimensions, cyber security, and AIS efficiency, but it does not establish causality due to the cross-sectional nature of the study design.Future research could consider this issue.Finaly, future research could consider employing alternative sampling techniques, ensuring a higher response rate, and incorporating objective measures or observations to complement self-reported data.Additionally, longitudinal studies or experimental designs could be employed to establish causal relationships between artificial intelligence dimensions, cyber security, and AIS efficiency.