XAI Based Intelligent Systems for Society 5.0

XAI Based Intelligent Systems for Society 5.0 focuses on the development and analysis of Explainable Artificial Intelligence (XAI)-based models and intelligent systems that can be utilized for Society 5.0—characterized by a knowledge intensive, data driven, and non-monetary society. The book delves into the issues of transparency, explainability, data fusion, and interpretability, which are significant for the development of a super smart society and are addressed through XAI-based models and techniques. XAI-based deep learning models, fuzzy and hybrid intelligent systems, expert systems, and intrinsic explainable models in the context of Society 5.0 are presented in detail.

The book also addresses—using XAI-based intelligent techniques—the privacy issues intrinsic in storing huge amounts of data or information in virtual space. The concept of Responsible AI, which is at the core of the future direction of XAI for Society 5.0, is also explored in this book. Finally, the application areas of XAI, including relevant case studies, are presented in the concluding chapter. This book serves as a valuable resource for graduate/post graduate students, academicians, analysts, computer scientists, engineers, researchers, professionals, and other personnel working in the area of artificial intelligence, machine learning, and intelligent systems, who are interested in creating a people-centric smart society.

• Defines the basic terminology and concepts surrounding explainability and related topics to bring coherence to the field
• Focuses on what techniques are available to improve explainability and how explainability can progress society
• Offers a broad range of topics, addressing multiple facets of XAI within the context of Society 5.0

• Language: English
• Publisher: Elsevier Science
• Release date: Nov 1, 2023
• ISBN: 9780323957847

Book preview

Section I

Paradigm shift and history of XAI

Outline

Chapter 1. Paradigm shift from AI to XAI of Society 5.0: Machine-centric to human-centric

Chapter 2. Towards explainable artificial intelligence: history, present scenarios, and future trends

Chapter 3. Society 5.0 and explainable artificial intelligence—implications

Chapter 4. Need for explainable artificial intelligence ethnic decision-making in society 5.0

Chapter 1: Paradigm shift from AI to XAI of Society 5.0: Machine-centric to human-centric

Marvin Paul Frank, and Ginu George     Department of Commerce, Christ (Deemed to Be University), Bengaluru, Karnataka, India

Abstract

Artificial intelligence (AI), the Internet of things (IoT), and robotics have gained significant momentum to meet expectations in many applications. Data management has become a tedious job as businesses grow. The interruption of AI in business functions and a growing web-based service economy in the last decade have led the IoT to grow faster, reducing the tedious job. Timely interruption of eXplainable artificial intelligence (XAI) reduces the technical complexities. On the one hand, the AI of Industry 4.0 promises the easiness of business functions. On the other hand, XAI of Society 5.0 tends to ease people's social life. This chapter ascertains the impact of AI on significant business functions and tries to bring out challenges AI faces and ethical values that must be considered in business functions. This chapter also tries to shed some light on the evolution of XAI of Society 5.0 and reasons for the shift from AI to XAI or machine-centric to human-centric and concludes by highlighting the future of XAI.

Keywords

Artificial intelligence; eXplainable AI; Human-centric; Machine-centric

1. Introduction

Artificial intelligence (AI) is the ability of machines to transcend human knowledge like human intellectual capacity, reasoning power, and learning experience from experience. During the computer era, AI and the Internet of Things (IoT) are widely used in all spheres of human life and the development of the corporate world. Digital technology is used to replicate human work. Today, technology is so advanced that it not only follows human commands but also commands machines to do tasks. The industrial revolution has enhanced industrial works by introducing and adopting wireless sensors, sensor-capable intelligent robots, and edge computing to accelerate the productivity and efficiency of work. The industrial revolution has generated many software and mobile applications for quicker communication. The following section briefly explains the concept of AI and IoT, its evolution, the role of AI in business functions, and ethical considerations in business functions. It also introduces the idea of eXplainable Artificial Intelligence (XAI), its evolution, principles, and the shift from machine-centric to human-centric, challenges of explainable AI, and finally, explains the future scope for XAI.

1.1. Concept of artificial intelligence

The term AI was first used in 1956 by cognitive scientist Marvin Minsky who was very optimistic about the technology in the future. John McCarthy, the father of AI, defines AI as the science and engineering of making smart machines, brilliant computer programs (Jain, 2019). AI consists of reasoning, programming, artificial life, belief revision, data mining, distributed AI, expert systems, genetic algorithms, systems, knowledge representation, machine learning, natural language understanding, neural networks, theorem proving, constraint satisfaction, and theory of computation (Cioffi et al., 2020). AI is defined on four categories firstly that system thinks like a human person, in a sense Haugeland defines as the exciting new effort to make computer think … machines with minds, in the full literal sense.Secondly that thinks rationally, in a sense Chamiak and McDermott define as the study of mental faculties through the use of computational models. Thirdly, a system that acts like humans in a sense Kurzweil defines AI as the art of creating machines that perform functions that require intelligence when performed by people. Fourthly, a system that acts rationally, in a sense Poole and his companions define AI as computational intelligence is the study of the design of intelligence agents (Introduction — mathematics and applications of machine learning, n.d.). Hence, AI works as a combination of thinking and acting like humans rationally. In short, we can say that AI is a science that aims to replicate aspects of human intelligence, such as learning, reasoning, perceiving, critical thinking, etc., using computer programs guided by logic. Saleh (Saleh, 2019) refers to AI as robots similar to humans in problem-solving and language understanding. Robots are programmed per the user's need, interact with the user through sensors, and are mostly autonomous. Even though AI has enormous utility and benefits, it has loopholes. Wang et al. (2019) comments that we cannot accurately define AI because a good description contains realistic aspects, but AI is based on a hypothesis. The purpose of AI does not provide clear instructions. But AI simulates the processes of the Intelligence of users by mobile devices and computer systems. Applications to AI consist of professional designs, natural language processing, voice recognition, and computer visuals.

1.2. Evolution of artificial intelligence and IoT in business

AI can be understood as the ability of a machine to learn from experience and adjust to new inputs given by human persons. The word AI was first introduced in 1950, and a lot of addition and deletion was made before making it final. There was high technology usage at a particular stage when big data technologies emerged. With the emergence of AI, there was much improvement in the storage capacities and super-fast speed in the data processes. With this, AI found meaning and recognition in the corporate world. Some studies reported that many corporates well accepted AI-enabled systems and transformed their business to a greater extent (Wilson & Daugherty, 2018). According to the survey of technology by Gartner in 2018, AI ranked first on the list of strategic technology. One of the reasons to be at the top is its ability to predict the future and its cost-effectiveness. The survey of Gartner concluded that around 59% of the business organization are still gathering the data and processing it using AI strategy.

IoT has increased machine work efficiency and it can be described as the network of connected objects from sensors, software, and other technological devices or gadgets to exchange data from one object to another with the help of the Internet. IoT identifies the issues with the work and provides commands to sort out the issues and rectify them. Components of IoT such as sensors, gateway, actuator, and light intensity detectors impact business functioning effectively and efficiently. IoT assists in connecting multiple devices to the Internet, thereby facilitating human person with business functioning and vice versa (Laudon & Laudon, n.d.). The IoT bridges the gap between the physical world and its representation in communication and information systems. From the business point of view, the application of IoT in manufacturing, supply chain management, finance, logistics, and finance has reduced the gap between the physical world of business and its information system. The IoT simplified the entire business process and has developed various business models to simplify business processes. Execution Languages for Web Services (WS-BPEL) is one of the processes which facilitate an interoperable integration model and integration model to support business transactions (Laudon & Laudon, n.d.). Smart Objects is another AI object part of IoT. Smart Objects are equipped with positioning and interactive technologies integrated into a communication network. Business processes use smart items in their functioning. Three essential smart items are radio-frequency identification (RFID), barcodes, and sensor networks. RFID device tracks the product's location in transportation, which makes the marketer and the customer follow the product quickly. A barcode is attached to the product, and the content of the barcode is detected optically. The reader of the barcode admits the printed data in the barcode and provides the data to the information system, which communicates the actual description of the product. Sensor networks are associated with RFID in the business process. The sensor networks can sense the business environment, like temperature, heat, spoilage, damage, and overall condition of the product. The sensed data are transmitted into the information system (Decker et al., 2008).

1.3. Impact of artificial intelligence on business functions

Industrial automation and robotics have existed long back and have made many impacts, but we will see a drastic change soon (Torresen, 2018). AI and automation have created heaps of advancement in recent decades. AI and automation will influence various business and service sectors (Nadimpalli, 2017). Tjoa and Guan (2015) in their research said that they have seen enormous AI and ML growth. Through his research, they had observed that some sectors require meticulous implementation of AI, especially in the medical industry, finance sectors, and business entities, which need high transparency and accountability. In recent years, many companies have undergone a significant change due to growth in technology. The technological increase transforms business models and functions, leading to revenue growth. Kunwar (2019) in his research found that as the fourth industrial revolution emerged in the 21st century, so approached AI. Around 2000 new businesses worldwide now have AI as a centerpiece of their action plan. With sophisticated data collection, analysis, and forecast mechanisms, AI technology shapes the company's daily operations to a greater extent and productivity. Active learning is the most reliable source to gather information and data with little cost and accelerated accuracy. Many research scholars have studied active learning strategies in different fields like video retrieval, speech recognition and classification, and text classification and concluded that these applications attract active learning.

Along with active learning, multilabel learning is also more prevalent (Hoi et al., 2006). Multilabel learning is one of the essential elements in today's corporate world. With the advancement in the corporate world and diversity among various business functions in the real world, multilabel learning has attracted wide attention in ML. Traditionally, many business firms held on to single-label learning (Qi et al., 2022; Zhang et al., 2018). Multilabeling is something other than accomplishing many tasks simultaneously. Securing data, image recognition, speech recognition, and tracking are done simultaneously and more effectively in business functions. Multilabeling active learning uses annotation, where information is labeled with images, audio, video, and text. These annotations are created based on assumptions made by experts. Today, most crowdsourcing platforms like Flipkart, Amazon, and other online business operators use it extensively and effectively (Wu et al., 2021). However, one of the issues regarding multilabeling is the cost and accuracy of the output. Since IoT is a recent trend in current business, yet to take the finest, the accuracy is only a few percent. Every advancement or technology costs a lot in the beginning. Similarly, much research must be done to make it readily available for every business unit at an affordable price (Wu et al., 2021).

1.3.1. Artificial intelligence in manufacturing

Currently, the manufacturing sector is entering into innovation and AI-driven production. Today, most manufacturing processes are integrated with robotics, IoT, automation, and machine-oriented manufacturing. The industry is moving toward intelligent manufacturing (Monostori et al., 2016). The ultimate goal of every manufacturing sector is to manufacture cost-effective, quality-based products. However, due to continuous technological changes, meeting such objectives has become increasingly complex and challenging (Sharp et al., 2018). The operation of AI in intelligent manufacturing has been built upon a series of new technologies over many years. AI has assisted the manufacturing process, increasing the production rate, improving quality and flexibility, reducing cost, and providing safety to a certain extent (Wang et al., 2019). Today, humans, known as cobots, do machine-assisted manufacturing work. Cobots are called collaborative or cooperative robots, assisted by IoT developed for humans safely. Cobots are comparatively lightweight and small, making production sectors very handy. Cobots will facilitate production safer and more conducive and provide better work atmospheres. AI can give safety to several dangerous jobs that unremarkably result in work injuries, leaving employees with minimum challenging work and allowing them to toil on complicated pieces freed from complications and accidents. ML algorithms can overcome several challenges and difficulties that arise once mistreatment of robotics and automation in the manufacturing sector, corresponding to when automation is programmed to finish precise work and cannot respond to unwelcomed circumstances. Thus, ML gathers the information, analyzes them, and detects unique designs (Zakharova et al., 2021). Another essential aspect of manufacturing is the manufacturing control system, which is one of the systems in production activity that assists in decision-making and controlling the entire activities of a factory to improve production efficiency and quality. AI is used to decide how much quantity to be produced. The time is set to complete the production. AI can also automatically evaluate the machinery types and predict maintenance needs. Thus, manufacturers can avoid major breakdowns and save costs (Çınar et al., 2021). Production throughput and bettered qualities are two essential components of the production department. AI with IoT provides 360-degree visibility across the machinery. Thus, manufacturers can reduce production costs and increase quality (Çınar et al., 2021). IoT assists the production department in planning, controlling, integrating, analyzing, and optimizing processes through networking among various machines, systems, human persons, and devices. This networking generates data and coordinates with each other, improving productivity, enhancing customer experience, and making the supply chain smooth. Control the production quality, manage the inventory, and predict maintenance during system breakdowns (Jiang et al., 2022).

1.3.2. Artificial intelligence in human resource management

Human resources (HR) has evolved steadily due to economic processes. Therefore, the immense developments in information technology have helped it cross numerous obstacles confining it to be treated as a mere body role within the organization. This growth in the HR functions is mapped through the constant shifts in its focus. Analyzing HR practices' impact on staff throughout the first phase became organization-centric. Phase 2 was once the HR swollen their scope to evaluating complete HR systems over the best HR practices. The ultimate phase saw a paradigm shift in HR when it was captive from HR to SHRM due to the belief that only HR might successfully align the organization's goals to the purposes of the personnel. Currently, the human resource management of every organization is motivated by the digital revolution, particularly AI along with the role of IoT (Amla & Malhotra, 2017). The impact of AI, particularly IoT, in HR can be seen within the type of E-HRM, wherever HR is more of a platform than a person.

Mittal (n.d.) comments on the requirement of AI in HR, and the world is completely turning out to be digitalized. Therefore, human resources must be completely up-to-date and up-skilled. Human intelligence must be on par with AI, if not better. HRM uses AI to screen, interview, and highlight new applicants' talent. Through the help of AI, the manager can examine the candidate's aptitude and do the best of the appointments. Though AI and human resources may sound like a trope in that individuals feel their purpose is to exchange them, the reality is different. Computing alludes to the technological innovation that permits machines to do tasks that may ordinarily need humans attributable to their intelligence. AI enhances human intelligence by relieving staff from indulging in automated tasks, forcing them to reinforce their data and skills to create much price for the organization. Human resource functions in corporations worldwide have adopted AI into their organizations, realizing its limitless potential and applications. The number of procedures being integrated with AI is increasing gradually. We can expect greater AI integration into other HR functions, such as automation in recruitment, induction, onboarding, and performance management. Kapoor (2019) highlighted in his research that many business entities adopt AI in human resource management because it is convenient to handle recruitment and selection. Even the study of Tambe et al. (2019) assured that most companies adopt AI to increase the speed of hiring a candidate and have an unbiased performance appraisal process and cloud-based HR system.

Many organizations use AI in HR, like machine learning, Automation, and chatbot, to reduce lengthy processes (Amla & Malhotra, 2017). AI is used in the process of recruitment. The study highlighted that 40% of companies and industries use AI in recruitment. AI is used to screen the applications, conduct interviews, and identify the required skills and talents of the candidates. Automated responding machines are used to solve queries of the information seeker. The AI-powered system is thoroughly used to recruit new employees for the business organization. It is used to receive and respond to an application, reduce the limited selection, train employees, track and guide the recruited, and effectively develop an organizational culture (Votto et al., 2021).

Resume screening is one of the essential tasks in HRM. The application of software and automation in recruitment is made for numerous reasons. One reason is that manually screening resumes consumes much time. Secondly, the majority of the applications received need to be qualified. Thirdly, to avoid the chances of being biased while screening applications. Human resource management uses the screening software system within the resume screening process. There are designed applicant–tracking—systems (ATS) that learn the task and then realize what qualified applicants appear based on earlier appointment decisions. Merging ML for information details is a game-changer for assessing the applicant through automation. AI fits into the recruitment process method within the screening stage. An intelligent screening process uses the information learned concerning applicants' skills, experiences, and other qualifications and credentials to screen and rank them automatically. AI is used in screening and in the process of interviews, where AI conducts speech pattern exams. Amy and Clara's tools schedule interviews and working meetings (Amla & Malhotra, 2017).

Selection is another stage in hiring a candidate for a job. Within a short period, AI identifies the right candidate with the required skills. In this process, AI can avoid favoritism and maintain transparency in recruitment (Rathi, 2018). The HR manager's role is to develop a healthy relationship with employees in the organization. Today's workforce in each industry is so vast and diverse that it is impossible to handle manually. The workforce is monitored constantly and observed minutely, enhancing safety and comfort. IoT simplifies everything within no time. IoT monitors employee growth, health, productivity, and comfort.

1.3.3. Artificial intelligence in marketing

The fundamental marketing objective is to provide suitable goods in the correct quantity and quality at the right time to the right person at the right place for the right price. Today, fulfilling this marketing objective is easy due to technological advancement (Decker et al., 2008). The widely used technologies in marketing are Automation, IoT, AI, and robotics. These advances, particularly AI and the IoT, will likely change business strategies in the upcoming years (Davenport et al., 2020). Using information, the system identifies traces location and monitors real-time responsiveness and optimization to achieve this objective. The tracking system detects the product in transportation, monitors the product conditions, and ensures the quantity and quality of the product (Decker et al., 2008). Some researchers envisaged AI immensely impacting forthcoming marketing due to IoT data analysis. Automation and AI in marketing form choices supported by data analysis and following patterns and trends, affecting marketing function. Thus, AI and IoT in marketing eliminate the risk of human error (Xie, 2019). The modern era is a customer-driven marketing era; along with the simplicity, more complexities in decision-making are constantly increasing. Customer-driven marketing includes understanding the needs and wants of the consumer and satisfying them. Changing customer behavior is significant in forming the most straightforward marketing decisions. AI and IoT are reshaping the features of marketing. It assists in reshaping pricing and sales and provides accuracy in research and development (Soni et al., 2019). Marketers spend lakhs of rupees on unfruitful advertising and marketing strategies, but through IoT, the cost and tedious work are waived with one click. IoT provides a solid technological basis for Marketing. Marketing uses wireless sensor networks and collects massive amounts of users' information through location-based social information (LBSNs). They gather location-based information like taste, preference, check-in location, check-in time, category, and feedback, thus, making the industry easy to meet customers' needs and serve better (Sharp et al., 2018).

IoT analyses customer behavior and reveals the pattern of product and service users, which helps marketers forecast demand and customer behavior. Thus, with the help of the IoT, marketers make strategies. IoT and AI try to enhance the customer's experience and daily dealings. It tries to bridge the gap between consumers and marketers (Soni et al., 2019). IoT in marketing solutions includes image recognition, chatbots, individual assistants like Google Assistant and Microsoft's Cortana, and dynamic pricing on E-Commerce sites. The companies that have widely used AI technologies are Google, Microsoft, Facebook, and Apple. AI and IoT are used in the service of Google Translator, Google Street View, Bing Search, Cortana Virtual assistance, and Siri. AI is usually utilized where speediness plays a dynamic role. With the help of speed, marketers can communicate with their customers effectively. Thus, marketers become prompter and alert to meet the demands of consumers (Soni et al., 2019).

1.3.4. Artificial intelligence in finance

The finance function refers to an organization's accounting section and financial processes. The finance function plays a vital role in an organization and has a strong relationship with its business entity (Waal, 2007). They were digitalizing through AI and IoT to transit the paper to a paperless-driven business transaction (Mittal, n.d.) and the use of AI in finance tend to shape and reshape cumbersome transactional and analytical routine financial processes and activities. IoT enhances and supports internal and external communication, protects information assets, and improves business decision-making (Kapoor, 2019). Digitalization has opened new opportunities for scholars to research, collaborate, and cooperate with the business industry by providing them with new digital enhancements (Tambe et al., 2019). According to Plaschke et al. (2018), around 42% of digital technologies can automate finance function activities. Automation in finance makes one rethink the organization's traditional tasks (Lambert & Sponem, 2011). Therefore, conventional finance functions must be converted into AI-powered and IoT-based. The role of the AI-based finance function is to provide value to the traditional finance function of the business. The AI-based finance function provider adds responsibilities to business analytics as an information expert and provides strategic directions using AI data analytics. Many researchers compared the traditional-based finance function and data-driven or AI-driven finance function. Their studies concluded that the AI-based finance function takes advantage of a paperless business environment which is faster and more cost-effective (Mittal, n.d.). The AI-based finance function is a trend that has enhanced finance methods and techniques. Changes in the management of finance functions are driven by the emergence of AI, such as IoT and cloud-based technologies, and by working faster, more thoughtfully, optimizing time, and sharing knowledge, data, and capabilities to reduce financial risks. Currently, in the finance function, a few well-known core processes are used to speed up the working capabilities through AI, such as procure-to-pay (P2P), order-to-cash (O2C), and report-to-record (R2R) enabled through AI technology and driven automation process. These AI-driven technologies focus on knowledge-based finance functions such as assessing future risk, external reporting, and forecasting through machine learning and analytics. AI is a strategic partner and creates extra value and worth in the organization's decision-making (Kapoor, 2019). Magnuson (2020) the author of research on Artificial Financial Intelligence, highlighted that P2P or R2R are core documents of finance function where a large volume of data entry and documentation is done, and many organizations use them. Therefore, AI, IoT, and automation must be tackled systematically and carefully. Some of the present automations in business functions are front-end imaging or optical character recognition (OCR), which recognizes the image and text; invoice workflow automation, which assists in accounts payable; and E-invoicing, which helps in the exchange of electronic documents between counterparties (Kapoor, 2019). IoT provides 360 degrees of customer view to enhance financial security and safety. It detects fraud and notifies the authority. IoT can even understand customers' needs and develop various auto-generated suggestions per their budget. In this way, IoT provides the best financial experience to the customers and is an easy task for the finance department of any industry (Decker et al., 2008). However, the immediate difficulty of converting to AI-powered is the need for digital skills among the employees of the finance function. Few researchers believe that the finance function has the available qualities to adapt to the changing digital environment (Richins et al., 2017). Mohamed and Lachine (Mohamed & Lashine, 2003) highlighted that finance functions of the modern world need necessary AI-related skills in the changing world business scenario. The finance function manages the financial resource of the business (Ricciardi & Simon, 2000). A significant role in finance functions is planning, budgeting, reporting accounting statements, internal control, compliances, operating risk, and risk management; today, it has done by the IoT (Cooper & Dart, 2013). The emergence of AI and IoT in the finance function has extended its function by adding some new features such as process optimization, change leadership, and performance management (Burns & Baldvinsdottir, 2007). Adding to this, some research scholars added yet few more features to the finance function, such as strategic planning, risk assessment, risk management, strategic control of finance, and forecasting financial growth and stability (Graham et al., 2012; Verstegen et al., 2007). Along with all these effective technologies for finance functions, some studies have discussed their drawbacks and dangers. According to Lyon (2018), automation is programmed to follow the rules and cannot be considered an intelligent machine compared to human intellect. He admitted that empathy is required toward job seekers less skilled to work in the current scenario; their mindset must be changed and encouraged to overcome the fear of change.

1.4. Ethical consideration of AI

Every new technology faces one or the other types of concern and issue. Ethical AI is the branch of technology that looks into the ethical aspects centered on new technology. The fundamental purpose of ethical aspects in AI and IoT is to design unbiased systems, treating all equally and fairly. Many discussions and deliberations have examined the role of AI and IoT in recent years, especially in automation in business. Wilson and Daugherty (Wilson & Daugherty, 2018) ascertained that AI should be used only to augment human judgment, not as automation. AI must be used to increase the intention of humans rather than replace human thinking and contribution. Shabbir and Anwer (2015) urged designers and users to be more morally good and ethical while using AI-powered methods, whether for personal or good purposes. Addressing the growth of AI, the English theoretical physicist, cosmologist, and the director of research at the Center for Theoretical Cosmology at the University of Cambridge, Stephen Hawking, made a remarkable comment on the impact of AI upon humanity that the development of complete AI could spell the end of the human race. To support the viewpoint of Stephen Hawking, Bill Gates warned the human race to be worried about the threat and challenges posed by AI.

Therefore, focusing on a responsible AI system, the AI user must prioritize ethical values while running either a profit-making or nonprofit organization. AI is a core component of all modern computer systems and software. Every aspect and dealing of our life is based on computers and AI. AI will negatively and positively dominate human life in the future. In his study, Stahl et al. (2021) identified 39 ethical issues. Some of them are the cost of innovation, Injury to physical integrity, lack of trust and confidence, lack of authenticity and quality data, impact on job opportunities, lack of privacy and transparency, unfairness, bias, and discrimination. There is some threat to human dignity in some cases because we are human beings (Siau & Wang, 2020). One of the ethical considerations of excessive use of AI is collecting and analyzing data. Software companies collect data and transfer it to manufacturing companies, where there are possibilities for manipulating and exploiting potential data. This explosion of data, known as the data deluge, unlocks various legal questions regarding data privacy and safety (Bessis & Dobre, 2016; Opresnik & Taisch, 2015). Sometimes, the data and potential information collected and shared using AI can be traded off for money and influence. It has to be drafted and implemented strictly to prevent these regulations from protecting from being manipulated. Collecting extensive data on business entities may threaten the business entities themselves (Lynskey, 2014). Considering the ethical part of the marketing function, Prahalad and Ramaswamy (2004) communicated that AI has moved the consumer from isolated to connected, from unaware to informed, from passive to active. People can use AI, but sometimes as the younger generation becomes obsessed with AI, there can be a context where breakage of traditional consumer law.

As AI grows and starts imitating humans, one of the questions is whether AI is sued while making errors or mistakes and causes enormous losses. In the case of causing danger to a human person due to breakage in AI, should the owner or designer of the software, AI, or user of AI be held responsible? Thus, we can conclude that AI certainly has the potential to make the life of people comfortable and better. Still, adopting AI is only possible if it is based on some guidelines (Prahalad & Ramaswamy, 2004). As the world advances, doing work artificially has some issues regarding how it works. The process of working with and understanding AI was challenging. At this juncture, world tech came up with another variant of AI called XAI. XAI is a set of processes and techniques that allows the users to comprehend and trust the results and output created by machine learning algorithms. XAI defines a type of AI model and its expected impact. XAI aims to bring transparency to AI adoption by simplifying and visualizing its decisions. The ultimate purpose of XAI is to create a sense of trust among the user toward AI and build a human relationship with AI, or rather, building human intelligence with AI (Karran et al., 2022).

Based on these reviews mentioned above, the chapter section sheds some light on a more profound level. The basic concept and evolution ofXAI are explained along with four principles guiding XAI's guiding force. The chapter also tries to justify how AI shifts from machine-centric to human-centric. And, this chapter concludes by keeping open for further research by highlighting its challenges to develop further research. Knowing the implication of AI and the role of IoT in various fields along with ethical considerations, the study moves onto the concept of XAI, which is more oriented toward the growth of the human person rather than the machine.

1.4.1. Organization of chapter

Section 2 elaborates the concept of XAI of Society 5.0. Section 3 enlightens shift from machine-centric to human centric analogy, and Section 4 concludes the chapter with future scope.

2. Concept of XAI of Society 5.0

2.1. Evolution of XAI of Society 5.0

Society 5.0 is a social concept that originated in Japan recently. The vision of Society 5.0 is to resolve many global social and economic challenges. Society 5.0 is a human-centered society that balances economic development growth with solving social problems and issues. Society 5.0 is a cyber-physical system (CPS) that integrates cyberspace and physical space. Fukuyama (2018) defined society as Society 5.0 is similar to Industry 4.0, but Society 5.0 takes a step ahead, representing a smart and intelligent configuration, a data-driven economy and society, emphasizing human desires and skills. The proposal envisages a fusion of the actual (physical) world with cyberspace to gather more reliable and customized data effectively for enhanced problem solving and value generation (Jordan & Mitchell, 2015). The evolution of society has taken different forms in history. In the beginning, the earliest society was known as the searching society (Society 1.0), followed by the farming society or agricultural society (Society 2.0), and then followed by the industrial society (Society 3.0). Later, society took another form known as information society (Society 4.0), and finally, achieved society (Society 5.0) (Gerston,