1. Introduction
The use of artificial intelligence (AI) in smart cities, its effect on governance, decision-making, innovative discipline, and prospects of revolution have been a topic of discussion in debate and practice recent years [
1,
2,
3,
4,
5,
6]. Data generation utilizing AI is possible in government and private sectors exploring new approaches to understand our world. The availability of big data may be effective in optimum usage of resources while making informed decisions [
7]. Artificial intelligence and the Internet of Things [
8,
9] can positively influence smart decision-making [
10]. At present, AI is becoming a need for daily life and organizational procedures as technology has taken great dives in empowering AI advancement [
11]. AI contributes to smart cities’ decision-making because smart decision-making utilizes a systematic and organized approach to collect data and applies rational decision-making systems rather than using hit and miss, instinct, or generalizing from overall experience [
12].
“Smart cities” is a multidimensional notion and has been defined differently by numerous scholars. However, the compulsory prerequisite to being a smart city is to attain sustainable social, environmental, and economic development and improve the living standards of society by utilizing Information and Communication Technology (ICT) and AI [
13]. The technological aspect of a smart city in the decision-making process can be defined as “a technologically interconnected city” or the use of artificial intelligence with big data to accomplish the intelligence and efficiency in managing the city’s resources [
14]. A study about smart decision-making in smart cities using big data [
12] introduced a three-layer framework characterizing a smart city as an “instrumented, interconnected, and intelligence”. Smart cities in the implemented phase utilize artificial intelligence and IoT for data acquisition by using surveillance cameras, meters, and sensor-based systems for real-time data and from open data sources and social media for quick response. The data collected through AI, IoT, and other sources are integrated and then transformed into a piece of relative information in the “interconnected” phase to deliver better insights for smart decision-making. Finally, transformed information gathered through data is envisioned to understand the city’s demands, requirements, needs, and policies. Hence, it can contribute to well-informed and smart decision-making [
15,
16].
The role of artificial intelligence in the SDM process has long been discussed and acknowledged by many smart city scholars [
17,
18,
19,
20]. However, different factors can impact decision-making in smart cities. Several scholars [
21,
22] have recently highlighted that decision-making in smart cities is not affected by ICT only. However, the city managers should listen to the people and stakeholders of the society and include them in this process. Evidently, [
23] found how digital towns may be utilized to embed planning and decision-making and design codes into the city’s e-governance. Moreover, [
24] suggested several measures available to improve the energy efficiency of smart cities, and city managers must compensate for energy, environmental, social, and financial factors to make smart decisions. Although it is found that AI has a positive impact on smart decision-making in smart cities, we believe that other important factors mediate between these two variables.
Despite the extensive research outline explained above, our knowledge of the exact use of AI on the SDM process in smart cities is still limited and inconsistent in many ways. For instance, [
25] argued that transparency should be seen both in designing AI assistants and the decision-making process, ensuring more legitimacy in the public eye rather than in the process. The advancement of AI in the form of learning algorithms is beneficial in expanding our comprehension of how the elevated smart city operates, but it is challenging to see how such approaches could ever dominate decision-making in the near future [
2]. What anticipated is that AI will inform the SDM process in the same manner that various computer instruments serve as the foundation of planning support systems. Moreover, [
4] identified the challenges linked with the impact of artificial intelligence-based systems on smart decision-making. They proposed a set of suggestions for IS scholars and discussed the integration of AI support to replace humans in decision-making in particular. Jarrahi [
26] provided a more pragmatic and proactive perspective by highlighting the complementarity of humans and AI by examining how the strength of each one can be utilized in the decision-making process in organizations characterized by complexity, uncertainty, and equivocality. Hence, the involvement of humans is important, and AI assistance can be utilized as a part of the decision-making process. Another significant weakness, we believe, is the lack of a foundational theoretical framework linking AI and the smart decision-making process via the mediating role of social innovation (the design and implementation of new solutions).
In their study, [
27] proposed a research framework to simplify the interaction between technology and social innovation to develop timely, pre-emptive, and sustainable plans and strategies for decision-making. They have figured out how to use economic incentives to benefit society while also protecting the environment. Gibson-Graham and Roelvink [
28] also highlighted a framework about social innovation, explaining that it is concerned with relegated social groups and their involvement in social decision-making. Drawing on these frameworks, we propose that the relationship between artificial intelligence and smart decision-making is indirect, rather than direct, mediated through social innovation.
This research aims to see how artificial intelligence, with the help of social innovation, influences decision-making in smart cities. To the best of our knowledge, we uncovered this gap in existing work, and we believe that our research will contribute to the literature on artificial intelligence and smart cities. This research will help academics and government officials better understand the need for social innovation and how it influences the interaction between artificial intelligence and smart decision-making in smart governance systems.
The remainder of the research is structured as follows: The literature review and hypothesis building are explained in
Section 2 after the introduction in
Section 1.
Section 3 describes the research methodology, data gathering, and data analysis. Tables, figures, and diagrams are used in
Section 4 to present the findings. Finally, in
Section 5, concluding remarks, limitations, and research directions are discussed.
4. Results
Table 2 describes the outcomes of KMO for all three variables (AI as independent, SI as mediating, and SDM as dependent variable) is 0.531, which is greater than 0.001. This suggests that the data sample size utilized for this research was adequate. Further, the Chi-square result is 777.933 with a substantial significant level of 0.000, which is again satisfactory.
As explained in
Table 3 below, the reliability analysis for 11 items were used to determine artificial intelligence’s influence as an independent variable, smart decision-making as a dependent variable, and social innovation as a mediator factor. The 11 questions given below were dispersed as follows: 3 items were allocated to artificial intelligence, 4 objects were assigned to social innovation, and 4 to smart decision-making. We utilized reliability analysis to discover the reliability for each factor. The outcomes revealed that overall, Cronbach’s Alpha was 0.914 of a total of 11 items with a sample size of 437, which indicated that all questions used to measure all the three factors were reliable for this study. Moreover, factor loadings for each component were greater than 0.7, except that of one A2, which was found to be 0.584. However, several previous studies accepted factors with a value higher than 0.5, so we added it as a reliable component. Factor loading for each component greater than 0.5 means that all questions asked from participants and used to measure factors were reliable and valid for this research.
Table 2 describes the outcomes of KMO for all three variables (AI as independent, SI as mediating, and SDM as dependent variable) is 0.531, which is greater than 0.001. It suggests that the data sample size utilized for this research was adequate. Further, the Chi-square result is 777.933 with a substantial significant level of 0.000, which is again satisfactory.
Table 4 explains the correlation among different variables, data reliability, and descriptive statistics. The value of the mean for artificial intelligence was 3.855 (SD = 0.804), indicating that respondents showed that they agreed to respond to the usage of AI in smart cities for decision-making and for social innovation 3.872 (SD = 0.754), showing that most respondents believed in social innovation for decision-making using big data, and for smart decision-making, the value was 1.266 (SD = 3.302), indicating that respondents were agreed strongly with smart decision-making in smart cities with SI and AI. The correlation between artificial intelligence and smart decision-making was (r = 0.811 **;
p < 0.01), which showed a significant positive relationship between both variables.
The correlation of artificial intelligence with social innovation was (0.679 **; p < 0.01), which revealed a similarly positive and significant relationship between AI and SI as assumed in the second hypothesis. Further, correlation (r = 0.414 **; p < 0.01) between social innovation and smart decision-making indicated a positive and significant relationship between independent and dependent variables, as anticipated in the third hypothesis. The Sobel test was conducted to test the mediating role of social innovation between the relationship of artificial intelligence and smart decision-making. Results from the Sobel Test given below explained a mediation association between independent and dependent variables.
Table 5 reveals a hierarchal multiple regression analysis to examine our research hypothesis, which stated that social innovation mediates artificial intelligence and smart decision-making. Regarding model 1, referring to the direct relationship between artificial intelligence and smart decision-making, the value of B = 1.276, and the value of Beta = 0.811 with
p-value = 0.000, which implies a substantial and positive relationship between artificial intelligence and smart decision-making, supporting our first hypothesis. As for model 2, which applied multiple regression analysis to discover both artificial intelligence as an independent variable and social innovation as a mediator variable with smart decision-making as a dependent variable, the outcomes revealed that the value of B = 1.545, and the value of Beta = 0.981 with
p-value 0.000, as an indirect relationship between artificial intelligence and smart decision-making; on the other hand, the value of B = 0.422, and the value of Beta = 0.252 with
p-value 0.000 as mediation between social innovation and smart decision-making. The findings established a positive and substantial direct and indirect association between artificial intelligence and smart decision-making. Moreover, social innovation has a strong positive and significant mediating impact between artificial intelligence and smart decision-making; hence, hypotheses 2, 3, and 4 are significantly supported empirically.
Table 6 demonstrates the outcomes of the Sobel test to analyze the mediation analysis. The results reveal the direct association between artificial intelligence and smart decision-making,
p-value = 0.000. It suggested that there is a considerable and positive direct correlation between artificial intelligence and smart decision-making. Furthermore,
p-value is 0.000 as an indirect association between artificial intelligence and smart decision-making. Moreover, the results proved a significant and progressive direct and indirect affiliation between artificial intelligence and smart decision-making; social innovation has a substantial positive and significant mediating role between artificial intelligence and smart decision-making.
5. Discussions and Implications
5.1. Discussions
Because of sophisticated algorithms, big data, and greater storage and processing capacity, AI systems are augmented with integrated components of digital systems, dramatically influencing decision-making. As a result, there is a growing demand for researchers of IT system social science to comprehend and analyze the impact of artificial intelligence on decision-making and contribute to the practical success and theoretical advancement of AI applications [
4]. Furthermore, we must identify and examine the indirect elements that might influence the good or negative relationship between AI and decision-making. This study aims to meet this need by recognizing, evaluating, and researching an important component in this research field, namely social innovation. Several earlier researchers discovered a moderating effect of elements between AI and decision-making [
85] and the mediating influence of elements between two constructs [
86,
87]. We implied social innovation as mediating factor to investigate the relationship between artificial intelligence and smart decision-making.
This study looked specifically at social innovation as a mediating element that might influence big data’s smart decision-making process. The findings demonstrated a significant mediating role of social innovation between AI and SDM. First, we looked at the direct association between AI and SDM, which was signed between both variables. Then, we incorporated social innovation as a moderating variable between independent and independent variables. An analysis of the direct and indirect interactions between AI, SI, and SDM revealed a substantial direct relationship between AI and SDM and an entire mediating influence of SI between AI and SDM. Smart city managers can gather, convert, and transport data using surveillance cameras, environmental sensors, electronic billboards, traffic management systems, charging stations, Wi-Fi, and other devices. While there is a high degree of AI use in the city, there is a greater possibility that the city’s governance will be better and that choices for the public will be made wisely.
This research aims to determine if social innovation has a mediating effect between external construct AI and endogenous construct SDM. We demonstrated this link with empirical evidence that it exists and is highly supported when studied. The first hypothesis predicted a substantial relationship between artificial intelligence and smart decision-making, which was established correctly, and we discovered significant outcomes. We employed SPSS software and regression analysis to establish the strength and significance of these relationships, which were substantially and significantly positive. Local governments in cities employ gadgets and devices connected to the internet. They are influenced by the usage of big data acquired through sensors and other artificial intelligence sources, which is a good sign.
Further, the second hypothesis predicted a direct influence of artificial intelligence on social innovation. The third hypothesis implicated a direct positive relationship between social innovation and smart decision-making. We proved this with an empirical investigation; hence, we conclude that such relationships exist between independent and dependent variables. Finally, we determined with our experimental testimony that social innovation has a strong, substantial mediating impact between AI and SDM, as we anticipated in our fourth hypothesis.
5.2. Implications
The findings of this study have ramifications for local city managers and smart city governors. Because our research was done in the public and private sectors in Pakistan and South Korea, city managers and governors may benefit from it in various ways. Small or rural communities may not benefit from this study due to a lack of resources, technology, social inclusion, political power, and other variables, but larger cities in these nations may benefit. The major goal of this study is smart decision-making; thus, we have emphasized the aspects that contribute positively and substantially so that local government managers should keep these factors in mind when developing public policies and choices. Sensors are becoming essential types of data collecting equipment that may be utilized for decision-making in smart cities for improved governance. Furthermore, local government can share such collected data with entrepreneurs, businesses, and industries, as well as for the prosperity of the society, and all relevant stakeholders, including such social innovators, should be involved in the decision-making process by local government, bearing in mind that such decisions will affect them directly or indirectly.
Because of various limitations, the findings and comments in this research should be interpreted with vigilance. Our first potential drawback is the very small sample size of 437 survey questions. Although there have been multiple earlier studies with smaller sample sizes than ours, we feel that the outcomes may be different with a larger sample size. Next, while we worked hard to avoid any social or nationalist pressures on the participants, it is conceivable that some of the contenders sensed some implied communal or patriotism, which required them to respond to the questions in favor of their country because people of many nationalities wants to demonstrate their nation as being better. Lastly, this survey was done in both developed and emerging economies, and most respondents were educated and earned a middle-income. There is a chance that the results will change if the sample is drawn from various economies or drawn from a population with a low level of education and income.
6. Conclusions and Future Research
Artificial intelligence has proven to be beneficial in a variety of industries. As AI has grown in popularity because of big data, enhanced algorithms, and increased processing power and storage, AI systems are becoming an integrated component of digital systems and significantly influence smart decision-making. Consequently, there is a growing need for social science and information systems researchers to examine and comprehend the ramifications for decision-making and contribute to AI technologies’ academic growth and empirical success. This work intends to meet this requirement by analyzing and emphasizing the curative function of social innovation in the relationship of AI and smart decision-making, emphasizing the significant issues and opportunities for future studies. Four research hypotheses are presented, focused on the usage and influence of AI for decision-making, with social innovation acting as a mediator. Our multiple regression results using SPSS suggest that AI using big data generated from sensors significantly influences social innovation and smart decision-making in smart cities. Furthermore, it is statistically proven that social innovation plays a substantial and important mediating role in the interaction between AI and smart decision-making.
Although the hypotheses given in this study are primarily for study in AI for decision-making and social innovation mediation, they can also offer valuable recommendations for research on the application and effect of AI in general, the impact of AI on decision-making on different industries, and, most importantly, the analysis of these significant relationships in different contexts using interaction variables to achieve interesting outcomes.