Artificial intelligence enabled management

Artificial Intelligence Enabled Management argues that the economic problems facing academics, professionals, managers, governments, businesses and those at the bottom of the economic pyramid have a technical solution that relates to AI. Businesses in developing countries are using cutting-edge AI-based solutions to improve autonomous delivery of goods and services, implement automation of production and develop mobile apps for services and access to credit. By integrating data from websites, social media and conventional channels, companies are developing data management platforms, good business plans and creative business models. By increasing productivity, automating business processes, financial solutions and government services, AI can drive economic growth in these emerging economies. Public and private sectors can work together to find innovative solutions that simultaneously alleviate poverty and inequality and increase economic mobility and prosperity

1. of Contents

1. Preface

2. Acknowledgments

3. Part I: Introduction to Data Enabled Management

1. Karunarathnage Sajith Senaka Nuwansiri Karunarathna, NarayanageJayantha Dewasiri, Rubee Singh, Mananage Shanika HansiniRathnasiri Chapter 1 What Does Artificial Intelligence–Powered ChatGPTBring to Academia? A Review

1. 1.1 Introduction2. 1.2 ChatGPT3. 1.3 Methodology4. 1.4 Discussion

5. 1.5 Conclusions and Implications

2. Karthik Shivashankar, Venkat Bakthavatchaalam Chapter 2 EducationPolicies Through Data Driven Decision Making: Accelerating InclusiveEducation for People with Disabilities

1. 2.1 Introduction2. 2.2 Background

3. 2.3 The Status of Children with Disabilities4. 2.4 AI as a Tool for Inclusion: A Review5. 2.5 Proposed Model

6. 2.6 Research Methodology7. 2.7 Data Collection

8. 2.8 Analysis and Policy Making

9. 2.9 Barriers to Implementing this Model in Developing Countries10. 2.10 Future Works

11. 2.11 Conclusion

3. Wasswa Shafik Chapter 3 The Role of Artificial Intelligence in theEmerging Digital Economy Era

1. 3.1 Introduction

2. 3.2 Artificial Intelligence Types

3. 3.3 Applications of AI in the Digital Economic Era

4. 3.4 Benefits of Artificial Intelligence in the Digital Economic Era5. 3.5 Challenges of Artificial Intelligence in the Digital Economic Era6. 3.6 Conclusion

4. Sonia Rani Chapter 4 A Review of Machine Learning Methods forDiagnosis and Classification of Thyroid Disease

1. 4.1 Introduction

2. 4.2 Motivation of the Study3. 4.3 Literature Review4. 4.4 Conclusion

5. Trisiladevi C Nagavi, Sanjana K G., P Mahesha Chapter 5 A Questionand Answering System Using Natural Language Processing and DeepLearning

1. 5.1 Introduction2. 5.2 Aim

3. 5.3 Objectives4. 5.4 Introduction5. 5.5 Applications6. 5.6 Existing System7. 5.7 Proposed System8. 5.8 Literature Survey

9. 5.9 Design and Implementation10. 5.10 Experimental Results

11. 5.11 Conclusion and Future Scope

4. Part II: Role of AI and Big Data in Management Functions

1. Kamalesh Ravesangar, Hafinas Halid, Syaza Lyana Mahadzir, RubeeSingh Chapter 6 The Reinvention of HRM Practices Through ArtificialIntelligence: Opportunities and Challenges in the Digital World of Work

1. 6.1 Introduction

2. 6.2 The Reinvention of HRM Practices Through ArtificialIntelligence (AI)

3. 6.3 Usage of AI-Based Software in the Human Resource

4. 6.4 Benefits and Challenges in the Digital World of WorkAssociated with AI in HRM

5. 6.5 Conclusion

2. Esther Jyothi Veerapaneni, Sampreeth Chowdary N Chapter 7 Challenges and Artificial Intelligence–Centered Defensive Strategies forAuthentication in Online Banking

1. 7.1 Introduction2. 7.2 Literature Survey

3. 7.3 Framework of Artificial Intelligence–Centred DefensiveStrategies for Authentication in Online Banking

4. 7.4 Result Analysis5. 7.5 Conclusion

3. Kamalpreet Kaur Paposa, Pardeep Bawa Sharma, ShubhamSuman Chapter 8 Catalyzing Human Potential: The Crucial Role of AI inModern HR Management

8. 8.8 Conclusion

4. Aneesya Panicker, Avnish Sharma Chapter 9 Exploring How ArtificialIntelligence is Changing the HRM Landscape: Refuting the Fiction withReality!

1. 9.1 Introduction

2. 9.2 Conceptualizing Artificial Intelligence

3. 9.3 Conceptualizing Technology Enabled Human ResourceManagement

4. 9.4 A Paradigm Changes in AI Integration Methods for HumanResource Management

1. 10.1 Introduction: Rationale and Motivation for the Book Chapter2. 10.2 Role of HR Chatbots in Employee Engagement

3. 10.3 Need and Significance of Chatbots in Employee Engagement4. 10.4 Challenges in Employee Engagement Kindle HR Teams

Toward HR Chatbots5. 10.5 Chatbots

6. 10.6 The Review of Literature

7. 10.7 Conversational Behavior Using Chatbots8. 10.8 Scope and Objectives of the Study9. 10.9 Research Design

10. 10.10 Findings and Recommendations

11. 10.11 Conclusion and Scope for Future Study5. Part III: Application of AI in Different Sectors

1. Hilmi Yüksel Chapter 11 An Empirical Analysis of Artificial IntelligenceApplications of Manufacturing Companies in Turkey

1. 11.1 Introduction

2. 11.2 Artificial Intelligence in Manufacturing3. 11.3 Research Objectives and Methodology4. 11.4 Discussion

5. 11.5 Conclusion

2. Rubee Singh, Kishore Kumar, Shahbaz Khan Chapter 12 AComprehensive View of Artificial Intelligence (AI)–Based Technologies forSustainable Development Goals (SDGs)

1. 12.1 Introduction2. 12.2 Emergence of AI

3. 12.3 AI and Innovative Sustainable Business Model4. 12.4 AI as an Enabler for Achieving SDGs

5. 12.5 Conclusion

3. Narayanage Jayantha Dewasiri, Dunusinghe G Dharmarathna, MrinaliniChoudhary Chapter 13 Leveraging Artificial Intelligence for Enhanced RiskManagement in Banking: A Systematic Literature Review

1. 13.1 Introduction2. 13.2 Methods

3. 13.3 Risk Management in Banking: An Overview4. 13.4 Utilising AI in Risk Management

5. 13.5 The Benefits and Challenges of AI in Risk Management6. 13.6 Challenges and Limitations of AI in Risk Management7. 13.7 Regulatory Landscape and Compliance Considerations

8. 13.8 Importance of Data Quality and Governance in AI-Driven RiskManagement

9. 13.9 Best Practices for Model Validation and Transparency

10. 13.10 Interpretation of Findings in the Context of the ResearchQuestion

11. 13.11 Implications of AI for Enhancing Risk Management inBanking

12. 13.12 Reflecting on Gaps in the Literature and Areas for FurtherExploration

13. 13.13 Emerging Trends and Innovations in AI for RiskManagement

14. 13.14 Conclusion

15. 13.15 Potential Areas for Future Research and Development

4. Soraya González-Mendes, Rocío González-Sanchez, Sara Muñoz Chapter 14 Exploring the Influence of Artificial Intelligence on theManagement of Hospitality and Tourism Sectors: A Bibliometric Overview

Alonso-1. 14.1 Introduction and Background2. 14.2 Methodology

3. 14.3 Results

4. 14.4 Thematic Organisation5. 14.5 Research Agenda6. 14.6 Conclusions

5. Garima Sainger Chapter 15 Artificial Intelligence in Healthcare Sector inIndia: Application, Challenges and a Way Forward

1. 15.1 Introduction

2. 15.2 Objective of the Study

3. 15.3 AI and the Healthcare Sector

4. 15.4 AI and the Indian Healthcare Sector

5. 15.5 Focus Areas of AI in the Healthcare Ecosystem6. 15.6 Potential Application of AI in Healthcare in India7. 15.7 Challenges of AI Application

8. 15.8 Recommendations and the Way Forward9. 15.9 Conclusions

6. Raghuveer Katragadda, Hari Babu Bathini, Sree Ram Atluri Chapter 16 Application of Artificial Intelligence and Machine-Learning Algorithms forForecasting Risk: The Case of the Indian Stock Market

1. 16.1 Introduction2. 16.2 Literature Review3. 16.3 Research Methodology4. 16.4 Data Analysis

5. 16.5 Conclusion6. About the Editors

Purpose: This chapter aims to provide a comprehensive picture of the adoption of

ChatGPT in academia, primarily focusing on the dark side of this emerging technology.This confession on ChatGPT relates to the four major stakeholders in academia:learners, teachers, researchers, and libraries.

Methodology: A narrative review was conducted based on extant literature on

ChatGPT to provide a comprehensive overview of the negativities carried by the tool toacademia.

Findings: Academia is one of the severely affected fields by ChatGPT The paper

reveals that all the critical stakeholders of academia, including students, teachers,researchers, and library staff, have been challenged by this emerging application.These roles are now topsy-turvy with the implementation of ChatGPT Though there arevarious rumblings, the authors were able to unveil critical drawbacks of the technologyin terms of high-tech plagiarism, tarnishing skills, hallucination, security and privacyissues, supplanting the jobs, promoting unfair teaching, the difficulty of using the tool,lack of accountability and lack of accuracy, and integrity of writing.

Practical Implications: The adoption of ChatGPT is inevitable in academia Therefore,

by identifying the consequences of ChatGPT, the learners, teachers, researchers, andlibraries can align their roles to mitigate the adverse impacts on academia Althoughacademia is in a hurry to adopt the changes the ChatGPT brings, our insights providesome avenues to re-think.

Keywords: ChatGPT, Education, Academia, Artificial Intelligence,

Chat platforms that can generate human-like responses have been able to dominateevery industry, changing their landscapes within a short period Siri search byApple and Amazon’s Alexa are good examples of this However, undoubtedly, ChatGPTstands as a global giant in the family of chatbots and has now reached our fingertipsfree of charge The power of ChatGPT has been able to penetrate through the veins oftoday’s generation However, professionals like scientists, researchers, andeducators are on the frontline of the user group (Frederick, 2023) This tech tool canpotentially facilitate each role of academia: teachers, students, researchers, andlibraries The speciality of ChatGPT is characterised by a personalised, interactive, andhuman-like nature that caters to an enhanced learning experience with autonomy – forinstance, providing customised recommendations and one-to-one interactive tasks andactivities that align with each student’s needs and learning goals (Biswas, 2023).

Further, students can use their capabilities more effectively at times if the app swiftlyproduces a passable response to a prompt or assignment (Anders, 2022) For teachers,ChatGPT can act as a virtual teaching assistant, providing immediate feedback on aspecific task, preparing question papers and quizzes, designing learning assessments,syllabi, and rubrics, etc Moreover, it can assess students’ assignments and other virtualtasks (Anders, 2022) Therefore, ChatGPT-based teaching is somewhat motivational(Ali et al., 2023) Some consider it more innovative and more efficient in academic workthan humans.

Similarly, ChatGPT would be beneficial in many areas of libraries For instance, they arefacilitating the patrons for 24/7 access to materials; handling simple and straightforwardinquiries of users and language translation; and facilitating services like referencecollection, development, and cataloguing, as well as selective dissemination ofinformation (Adetayo, 2023) Additionally, it is revolutionary that this tool has beenrecognised with the co-authorship of the research work by King and ChatGPT (2023).On the contrary, some claim ChatGPT is a destructive technology due to itscontroversial virtues It has enormous prospects to create risks in education andacademia with its extraordinary potential Elbanna and Armstrong (2023) argue that theadoption of ChatGPT in education and its consequences are still uncertain The noveltool possesses the potential to change every aspect of education, especially theconventional methods of learning, teaching, assessment, and evaluation Post-ChatGPT launch, educational institutions are doubtful whether this tool is accepted dueto the pros and cons offered, like enhanced learning from one side and academic

cheating on the other First, the platform challenges existing performance evaluationsystems because students can easily claim answers generated via ChatGPT as theirown Thus, this capacity challenged the foundation for maintaining and improvingprogram quality and students’ learning outcomes (Chaudhry et al., 2023) However,there is no evidence to say that the deliberate aim of the application is to utilise it as aproxy for academic writing However, it provides the room for artificial intelligence (AI)ingenuity Therefore, despite the benefits, students can abuse it (Benuyenah, 2023).However, without a holistic view of the tool’s possibilities and negativities, highereducation and academia players are doubtful about its applications (Sun et al., 2023).According to McKinsey (2022), the education sector is far behind in adopting ChatGPTrelative to the leading sectors like high tech, communications, financial services,automotive services, and energy This can be attributed to the adversities associatedwith ChatGPT in education Countries like Italy have banned ChatGPT, and theCanadian government has started investigating OpenAI due to the black-box nature ofthe technology Due to the ethical issues related to authorship, journals

like Science have banned the usage of large language models (Hosseini et al., 2023).The overwhelming sensation towards ChatGPT will be more balanced and reflectivewith the gradual recognition of its faults and limitations with time It has been asignificant concern to figure out how to deal with ChatGPT Benuyenah (2023) pointedout that the effect of AI in higher education cannot be neglected or denied Therefore,educational institutes must embrace AI amidst all the pros and cons (Dewasiri et al.,2023a) Ali et al (2023) highlight that ChatGPT should be adopted for learning despitefears of its shortcomings and requires further detailed investigations Although theresearch on this technological tool is increasing, many avenues exist to focus on Giventhe nascency of ChatGPT as a field to research, this chapter aims to provide insightsinto issues and newly emerged challenges brought about by the tool for highereducation and academia.

The eyes of the world turned towards artificial intelligence (AI) after the release ofsoftware termed ChatGPT in 2022 by OpenAI ChatGPT is a significant LLM artificialintelligence program that has caused revolutionary changes in fields like media,schools, academia, and research environments It can be accessed easily by anyonewho has a device and internet connection This chatbot can generate human-likeresponses to the users’ input prompts ChatGPT’s advanced search engine, embeddedwith a large volume of trained data, enables it to create diverse replies like text, code,images, audio, and videos with different contents and lengths (Korzynski et al.,2023; Singh & Singh, 2023) A much more advanced version of GPT-3 was introducedin March 2023, known as GPT-4 Based on the data gathered from its predecessor, thenewer version can understand the meaning of significant texts and answer correctly,translate texts accurately, summarise contents and code, etc (Elbanna & Armstrong,2023) Though GPT-3.5 is a free version, GPT-4 is the paid version Also, now it has asubscriber service called “ChatGPT Plus.”

A narrative review was carried out to bring modest and novel insights using recentpapers published These papers cover the critical aspects of academia: learning,teaching, researching, and library services The article selection was based onthree criteria First, we used sources from widely accepted publishing houses likeScienceDirect, Taylor & Francis, Emerald Insight, Elsevier, Springer, Wiley, and GoogleScholar to ensure the reliability of the insights Secondly, we determined the cut-offpoint of the articles as 2015 to protect the novelty of insights Finally, we includedpapers related to the negativities of ChatGPT in education and academia These criteriaensure the relevancy of the contents.

Further, these hallucinations consist of undetectable logical and practical errors(Frederick, 2023) When researching, it is even worse that the tool can come up withentirely fabricated titles and authors (Burger et al., 2023) Therefore, students andresearchers are misled by the device, and it is up to librarians to make their users awareof it.

1.4.2 High-tech Plagiarism

High-tech plagiarism is a significant concern in assessments and research produced assignments can easily pass academic integrity tests from the widely usedtool in academia to validate authorship, Turnitin Further, devices like GPTZero and

ChatGPT-Copyleaks, which detect AI-generated content, must be improved to catch up with theAI-generated content (Chaudhry et al., 2023) Therefore, ChatGPT allows fordemocratising plagiarism and threatens academic integrity (Farrokhnia et al., 2023).The issues of originality and writing ethics, like plagiarism, also lead to the spread offake and double academic works Therefore, the challenge of distinguishing AI-generated content from human-generated ones is before the educators.

1.4.3 Lack of Accuracy and Integrity

Foundation laid on pre-trained and outdated data has brought accuracy and integrityissues This limited comprehension occurs due to the frame of the training data set andthe inability to catch up with human feelings and emotions (Adetayo, 2023) Cano et al.(2023) stated that inaccurate responses popped up from the AI model due to the needfor more resources and limited knowledge For instance, some terms of Englishvocabulary are not included in the chatbot platforms (Qasem et al., 2023) ChatGPT hasbeen trained with a vast amount of data from multiple disciplines; the accuracy and theintegrity of scientific writing are questioned because ChatGPT is embedded with pre-trained outdated data until 2021 (Ivanov & Soliman, 2023; Gao et al., 2022) Therefore,researchers must think twice when generating research ideas solely relying on AI Thesubstantial developments of realms after this cut-off date may not be included in theoutputs of ChatGPT (Burger et al., 2023) Therefore, these ideas are not insightful andnovel Limited knowledge is further confirmed by the fact that ChatGPT has only beentrained on open-access (OA) papers Therefore, most closed-access documentsprotected by paywalls but with valuable and updated content are usually not included inAI training (Burger et al., 2023) As many scholars have published non-OA journals,ChatGPT cannot fetch the original contents of such papers while citations and abstractsare available (Frederick, 2023) Therefore, the tool produces conclusions based onlimited data within the abstracts of non-OA journals.

Propagation of misinformation is further fuelled by predatory journals often published inopen access (Burger et al., 2023) As Frederick (2023) highlighted, ChatGPT containsdata from Google Scholar, which leads to the inherent limitations of junk papers andother fraudulent publications Thus, the integrity of literature surveys is threatened andcreates severe issues in academic writing Similarly, inaccurate responses to thequeries are generated during reference transactions (Adetayo, 2023) Therefore,librarians have to concentrate on the same.

1.4.4 Lack of Accountability

Two characteristics of the AI model harm accountability: first, content produced withouta source and credit to the original author (Qasem, 2023), and second, the fact thatusers need to learn about the rationale underlying the answers displayed by the deeplearning model, which is critical in scenarios that affect the community’s well-being(McKinsey, 2018).

Another argument directed towards ChatGPT is that it is solely text – a generatingmachine powered by AI Therefore, it is not accountable for the accuracy or relevancy ofthe information, which will result in some violations of the regulations that ban itsutilisation Though it is also conceivable that ChatGPT technology will becomeubiquitous before institutions have the time to amend their policies, there is an urgentneed for a practical approach that focuses on correcting the issues (Mhlanga, 2023).However, by today, there is no immunity to these matters Renowned publishers likeElsevier have now accepted ChatGPT as a co-author, but authors are responsible forthe results of the works In that case, are large language models accountable for theoutputs? Researchers must be thorough about these facts and implement necessaryprecautions while working with the tool (Frederick, 2023).

1.4.5 Unfairness in Teaching

The accessibility to ChatGPT discriminates against students (Qadir, 2022) Chat GPTprovides an advantage for the users over non-users As Cotton et al (2023) claimed,this is common in the assessment process, where students who use the toolmanufacture high-quality assignments and have an unfair advantage over studentswithout accessibility This is highly inequitable regarding summative assessments ofgrading the students On the other hand, users also have some drawbacks whereteachers do not measure the actual capacities of ChatGPT users Therefore, teacherscannot support the progress of the ChatGPT using students in formative-typeassessments Hence, teachers should have one single approach for all the students,not being in a hurry to adopt the technology Also, ethical guidelines are a must toenhance fairness and mitigate bias (Cotton et al., 2023) Further, ChatGPT does nothave much support for auditory learners who love hearing.

1.4.6 Skill Tarnishing

Another argument from scholars is that ChatGPT kills the higher-order cognitive skills ofstudents (Farrokhnia et al., 2023) Therefore, the expected skills that suit the future jobmarket are not catered to by ChatGPT-based learning (Chaudhry et al., 2023) Theresearch community is also keen on the influence of AI language models (Qasem,2023) Qasem (2023) provides evidence that ChatGPT can potentially increase theusage of ready-made materials instead of extensive reading and exploring the researchfacts This harmful habit of machine dependency among researchers and students killscreative and rational thinking.

1.4.7 Supplanting Jobs

According to Inamdar (2023), technology must enhance human knowledge rather thanreplace it However, ChatGPT can replace all three roles, namely teacher, researcher,and library worker, to some extent with its capacities Some educators are concernedthat AI can replace their part and reduce the need for their involvement (Elbanna &Armstrong, 2023) Writing and thinking with the aid of artificial intelligence and beingtaught by artificial intelligence are also significant concerns (Mohammadzadeh et al.,

2023) The teachers of higher educational institutions have become fans of ChatGPTdue to its convenience However, teachers have to consider the delivery quality whenutilising the tool Just banning its use might not work out in the long run Hence, itdemands adjusting the teaching styles amalgamated with chatbots To this end,educators have to change their teaching methods to group discussions, presentationson recent events, case studies, innovative posters, and role-playing, like things thatinvolve higher-order thinking skills and independent learning.

Moreover, teachers must focus more on teaching skills and attitudes (Dewasiri et al.,2023b) Therefore, ChatGPT teachers must transform their teaching styles fromdelivering fact-full lectures to innovative methods, which signals that the teachers whorely highly on traditional methods will be replaced by this tool Furthermore, ChatGPTcan supplant the jobs of library workers because repeated functions like cataloguingand indexing are quickly done using ChatGPT Therefore, the device may soon replacecuration and cataloguing personnel (Inamdar, 2023) Even this can supersede theactivities of journalists, playwrights, and educators to some extent Further, the roles ofresearchers, like literature searches, computer code writing, and data processing, willbe dominated by ChatGPT soon (Frederick, 2023).

1.4.8 Security and Privacy Issues

Privacy concerns and unreliable and fake information have been an overly complexissue brought about by the AI language models that is expected to happen extensively(Qasem, 2023) These privacy and security issues result in legal actions for intellectualproperty and copyright Legal matters like copyright problems, incorrect citations, andcybersecurity risks are frequent in education (Sallam, 2023) Also, it can causeprejudices and systemic biases in the trained data, neglecting intellectual property rightsand ignoring the values and norms of companies (McKinsey, 2022; Inamdar, 2023).Also, ChatGPT does not reference the derived conclusions Users need help finding thetitles, authorship, publisher, date, or place of publication (McKinsey, 2022; Frederick,2023) Users have no idea about the nature of the data collected, the sources of thedata, and the protection of privacy and security (Frederick, 2023; Elbanna & Armstrong,2023) Although the referrals based on ChatGPT’s program instructions are fake, theyare likely credible These bogus citations tend to be misinterpreted as genuine by theusers unknowingly Once ChatGPT is adopted in libraries, it raises privacy concernssince it can collect and evaluate students’ and library patrons’ data, which may not besecure (Elbanna & Armstrong, 2023; Inamdar, 2023).

1.4.9 Difficulty in Adoption

Even though ChatGPT has captured an enormous portion of higher education, itcomprises technology-oriented barriers First, the outputs generated via the model havedramatic variations depending on the nature of the query, including specific words,syntax, and complexity There are mismatches of the generated answers, though thequestions and inquiries remain pretty similar (Qasem, 2023) The same question,

identical in meaning but with different synonyms, could manufacture eloquent output tozero output Therefore, learners must try several paraphrased queries to generate thebest possible answer.

Moreover, it claims the documentation of exact wordings of the AI prompts and all inputparameters (Burger et al., 2023) Also, libraries should be aware of optimising thequeries to produce ideal search results (Frederick, 2023) In addition, though ChatGPThas the potential to generate tailored responses, limitations could occur regarding thelength of the produced assignment and the response time For example, ChatGPT canwrite 100–300 words of relevant text for one prompt Therefore, many prompts shouldbe used to generate assignments with more word requirements (Ivanov & Soliman,2023).

1.5 Conclusions and Implications

The ChatGPT dilemma continues Though ChatGPT has increased the efficiency ofnumerous aspects of academia, there are still significant areas to focus on, especiallyfor the main parties involved in higher education: students, teachers, researchers, andlibrary workers To tackle the identified pitfalls, a collaborative effort not only from theusers but also from the developers is compulsory To deal with high-tech plagiarism,software developers should focus on more advanced AI-generated content detectionmodels that can differentiate from genuine content For this, NFT (non-fungible token)technology is another promising solution Since blockchain technology is embedded inNFTs, it has a higher level of security NFTs can make digital objects like images, texts,and videos non-fungible and unique, providing code that includes metadata(Mohammadzadeh et al., 2023) To enhance integrity and accuracy, AI models shouldbe fed with suitable models and frameworks (Inamdar, 2023) Developers should try todiscount sources of misinformation, avoid hallucinations, and minimise the difficulty ofworking with the software This can be achieved via much more advanced versions ofthe applications.

We are heading to a more advanced conceptual world where just remembering facts isnot expected by the learners Over-reliance or higher dependency discourages thematurity of critical skills Therefore, students are proposed to embrace the roles ofprompt engineer, fact checker, and editor rather than solely relying on results retrievedfrom ChatGPT verbatim Simultaneously, teachers can focus on enhancing thechildren’s creativity to fit the next decade They can improve 21st-century skills liketechnology, media, and information skills Since ChatGPT has created many avenues tofacilitate learning regarding course development and design, lesson planning,assessment, and evaluation, educators can align their roles accordingly For instance,they focus on the assignment process rather than the final text, teaching students to usethe tool constructively, protecting ethics, and highlighting limitations Cotton et al.(2023) suggest a few strategies to abide with ChatGPT in teaching The first strategy isto design assessments that demand skills like critical thinking, communication, andproblem-solving The second strategy is to ask students to develop proper citations andreferences, which may guide the student to seek more reliable sources of information

rather than sticking to just AI chatbots (Cotton et al., 2023) The third strategy is toprepare open-ended questions that encourage originality and creativity (Cotton et al.,2023).

Just banning ChatGPT would not benefit the progress of scholarly works either.Therefore, it is suggested that the technology be embraced with accountability,accuracy, and integrity To this end, disclosing the use of ChatGPT in the Introduction orMethods section, citations and references, and supplementary materials or appendiceswould be beneficial (Hosseini et al., 2023) Also, to minimise prejudices and ensure thatthe content is moral and just, libraries must be equipped with specialised training aboutsecurity and privacy (Inamdar, 2023) Libraries must eliminate the higher dependencyon AI since it hinders the personalisation of the offerings Efficient use of AI toolsrequires a general core understanding of AI and specific skills like Python Therefore,institutions can offer courses for students, teachers, researchers, and library staff ontechnology (Burger et al., 2023).

Adetayo, A.J (2023) Artificial intelligence chatbots in academic libraries: the rise of

ChatGPT Library Hi Tech News, 40(3) 18–21 0007 a, b, c

https://doi.org/10.1108/LHTN-01-2023-Ali, J.K.M., Shamsan, M.A.A., Hezam, T.A and Mohammed, A.A.Q (2023) Impact of

ChatGPT on learning motivation: teachers and students’ voices Journal of English

Studies in Arabia Felix, 2(1), 41–49 doi: 10.56540/jesaf.v2i1.51. a, b

Anders, B.A (2022) Why ChatGPT is such a big deal for education availableat: https://scalar.usc.edu/works/c2c-digital-magazine-fall-2022—winter-2023/why-

chatgpt-is-bigdeal-education a, b

Benuyenah, V (2023) Commentary: ChatGPT use in higher education assessment:Prospects and epistemic threats Journal of Research in Innovative Teaching &Learning, Vol 16 No 1, pp 134–135 https://doi.org/10.1108/JRIT-03-2023-097 a, b

Biswas, S (2023) Role of Chat GPT in Education Journal of ENT Surgery Research,

1(1), 1–3 Available at SSRN 4369981. →

Cano, M.Y., Venuti, F., & Martinez, H.R (2023) ChatGPT and AI Text Generators:Should Academia Adapt or Resist?, Harvard Business Publishing, availableat: https://hbsp.harvard.Edu/inspiring-minds/chatgpt-and-ai-text-generators-should-academia-adapt-or-resist →

Chaudhry, I S., Sarwary, S A M., Refae, G A E & Chabchoub, H (2023) Time torevisit existing student performance evaluation approach in the higher education sector

in a new era of ChatGPT – A case study Cogent Education, 10(1) 2210461,

DOI: 10.1080/2331186X.2023.2210461 a, b, c

Choi, J H., Hickman, K E., Monahan, A & Schwarcz, D B (2023) ChatGPT goes to

law school SSRN Electronic Journal, https://doi.org/10.2139/ssrn.4335905 →

Cotton, D.R., Cotton, P.A., & Shipway, J.R (2023) Chatting and cheating: ensuring

academic integrity in the era of ChatGPT Innovations in Education and Teaching

International, 1–12, doi: 10.1080/14703297.2023.2190148. a, b, c, d, e

Dewasiri, N.J., Karunarathne, K.S.S.N., Menon, S., Jayarathne, P.G.S.A, & Rathnasiri,M.S.H (2023a) Fusion of Artificial Intelligence and Blockchain in the Banking Industry:Current Application, Adoption, and Future Challenges In Saini, A and Garg, V (Ed.)Transformation for Sustainable Business and Management Practices: Exploring theSpectrum of Industry 5.0, Emerald Publishing Limited, Bingley, 293–307 https://doi.org/10.1108/978-1-80262-277-520231021 →

Dewasiri, N.J., Pigera, A.K.M., Karunarathne, K.S.S.N and Rathnasiri, M.S.H (2023b).Financial Services Employee Engagement and Attitude Toward Artificial Intelligence:Evidence From Sri Lanka In Saini, A and Garg, V (Ed.) Transformation for SustainableBusiness and Management Practices: Exploring the Spectrum of Industry 5.0, EmeraldPublishing Limited, Bingley, 231–245 https://doi.org/10.1108/978-1-80262-277-520231017 →

Hosseini, M., Resnik, D B, & Holmes, K (2023) The ethics of disclosing the use of

artificial intelligence tools in writing scholarly manuscripts Research

Ethics https://doi.org/10.1177/17470161231180449 a, b

Farrokhnia, M., Banihashem, S K., Noroozi, O & Wals, A (2023) A SWOT analysis of

ChatGPT: Implications for educational practice and research Innovations in Education

and Teaching International, 1–15 DOI: 10.1080/14703297.2023.2195846. a, b

Frederick, D.E (2023) ChatGPT: a viral data-driven disruption in the information

environment Library Hi Tech News, 40(3), 4–10 2023-0063 a, b, c, d, e, f, g, h, i

https://doi.org/10.1108/LHTN-04-Gao, C.A., Howard, F.M., Markov, N.S., Dyer, E.C., Ramesh, S., Luo, Y & Pearson,A.T (2022) Comparing scientific abstracts generated by ChatGPT to original abstractsusing an artificial intelligence output detector, plagiarism detector, and blinded human

reviewers bioRxiv Doi: 10.1101/2022.12.23.521610. →

Elbanna, S & Armstrong, L (2023) Exploring the integration of ChatGPT in education:adapting for the future Management & Sustainability: An Arab Review, Vol ahead-of-print No ahead-of-print https://doi.org/10.1108/MSAR-03-2023-0016 a, b, c, d, e, f, gINSEAD (2023) ChatGPT and the future of business education Availableat https://knowledge.insead.edu/leadership-organisations/chatgpt-and-future-business-education →

Inamdar, S (2023) Impact of artificial intelligence text generators (AITGs) on

libraries Library Hi Tech News, Vol print No

ahead-of-print https://doi.org/10.1108/LHTN-03-2023-0048 a, b, c, d, e, f

Ivanov, S., & Soliman, M (2023) Game of algorithms: ChatGPT implications for the

future of tourism education and research Journal of Tourism Futures, 9(2), 214–

221 https://doi.org/10.1108/JTF-02-2023-0038 a, b

King, M.R., & ChatGPT (2023) A conversation on artificial intelligence, chatbots, and

plagiarism in higher education Cellular and Molecular Bioengineering, 16(1), 1–2 →Korzynski, P., Mazurek, G., Altmann, A., Ejdys, J., Kazlauskaite, R., Paliszkiewicz, J.,Wach, K & Ziemba, E (2023) Generative artificial intelligence as a new context for

management theories: analysis of ChatGPT Central European Management Journal,

31(1), 3–13 https://doi.org/10.1108/CEMJ-02-2023-0091 →

McKinsey (2018) Artificial intelligence is a moving target Here’s how to take better aim.Available at: https://www.mckinsey.com/capabilities/quantumblack/our-insights/what-ai-can-and-cant-doyet-for-your-business →

McKinsey (2022) McKinsey technology trends outlook 2022 Availableat: https://www.mckinsey.Com/capabilities/McKinsey-digital/our-insights/the-top-trends-in-tech a, b, c

Mhlanga, D (2023) Open AI in education, the responsible and ethical use of ChatGPTtowards lifelong learning available at: https://ssrn.com/abstract54354422SSRN →Mohammadzadeh, Z., Ausloos, M, & Saeidnia, H.R (2023) ChatGPT: high-techplagiarism awaits academic publishing green light Non-fungible tokens (NFT) can be a

way out Library Hi Tech News, 40(7), 12–14 0067 a, b

https://doi.org/10.1108/LHTN-04-2023-Burger, B., Kanbach, D.K., Kraus, S., Breier, M., & Corvello, V (2023) On the use of

AI-based tools like ChatGPT to support management research European Journal of

Innovation Management, 26(7), 233–241 https://doi.org/10.1108/EJIM-02-2023-0156 a, b, c, d, e, f

Qasem, F (2023) ChatGPT in scientific and academic research: future fears and

reassurances Library Hi Tech News, 40(3), 30–32 2023-0043 a, b, c, d, e

https://doi.org/10.1108/LHTN-03-Qasem, F., Ghaleb, M., Mahdi, H.S., Al-Khateeb, A & Al Fadda, H (2023) Dialog

chatbot as an interactive online tool for enhancing ESP vocabulary learning Saudi

Journal of Language Studies, 3(2), 76–86 https://doi.org/10.1108/SJLS-10-2022-0072 →

Qadir, J (2022) Engineering Education in the Era of ChatGPT: Promise and Pitfalls ofEnerative AI for Education TechRxiv, Preprint, availableat: https://www.techrxiv.org/articles/preprint/Engineering_Education_in_the_Era_of_ChatGPT_Promise_and_Pitfalls_of_Generative_AI_for_Education/21789434 →

Sallam, M (2023) ChatGPT utility in healthcare education, research, and practice: a

systematic review of promising perspectives and valid concerns Healthcare, 11(6),

doi: 10.3390/healthcare11060887 →

Singh, H & Singh, A (2023) ChatGPT: systematic review, applications, and agenda for

multidisciplinary research Journal of Chinese Economic and Business Studies, 21(2),

This research explores the development of data-driven management practices andpolicies to address the educational needs of children with disabilities within an evolvingeconomic landscape Around 1.3 billion people globally live with disabilities, with 80% indeveloping nations The proposed model, targeting these marginalised groups,incorporates a data-driven approach segmented into three levels: school, regional, andnational In this work, initial non-intrusive data collection at the school level identifiesstudents’ needs based on demographics and unique requirements Regional bodiescollect similar data from the participating schools to define policies tailored to theparticular needs of the demography Centralised data from various regions informsnational-level analysis, allowing for evaluating teaching practices and leading to data-informed policies and procedures This interconnected system promotes thedevelopment of indigenous practices at school, regional, and national levels andensures adequate documentation and sharing of successful strategies Integrating AIwith wearable technologies, tailored content delivery, virtual classrooms, andempowering educators provides a multifaceted solution Collaborative efforts betweentechnology companies, educational institutions, and governments are vital forovercoming existing challenges This research illustrates AI’s potential in crafting aninclusive, personalised, and efficient learning environment for children with disabilities,particularly in emerging economies, which would be helpful for special needs schoolmanagement, policymakers and governmental bodies.

Keywords: Education Policies, Disability, Inclusive Education, Artificial Intelligence, Data Driven

Approach, Developing Countries,

In the contemporary world, education stands as a universal necessity, playing a criticalrole in the lives of children, regardless of their abilities (UNICEF, 2017; UNHCR, 2014).The importance of education is magnified for children with disabilities, as itempowers them to not only develop essential skills and knowledge but also fostersintegration into society Guided by Sustainable Development Goal 4 (SDG4), theemphasis on inclusivity and justice has never been more vital, resonatingwith UNESCO’s (2017) call for leaving no one behind Although inclusivityencompasses a wide array of attributes such as gender, age, ethnicity, and financialstatus (Antoninis et al., 2020), this research narrows its focus on the intersection ofchildren with disabilities, AI technology, and the context of emerging economies tocreate management practices.

The challenges faced by children with disabilities transcend geographical boundaries,and the situation is even more precarious in developing countries, where issues stemfrom various social, economic, and cultural barriers (Antoninis et al., 2020) Educationoffers a path for children with disabilities to overcome these obstacles, nurturing theirself-confidence and self-advocacy skills and protecting them from discrimination(Dyssegaard & Larsen, 2013; Hayes & Bulat, 2017; Kuper et al., 2018; Szumski et al.,2017).

Educational management practices and policies rely heavily on fragmented pockets ofevidence, needing a centralised repository conducive to developing policies informed bycomprehensive, large-scale metadata This research endeavours to address this gap bydelving into the possibilities and barriers surrounding collecting and analysing extensivedata on a large scale in an emerging economy The ultimate goal is to leverage thisdata to formulate evidence-based policies to enhance educational managementpractices The past decade has seen significant strides in adopting inclusive educationpolicies, increased teacher training, and AI-driven solutions to address uniquechallenges (Dyssegaard & Larsen, 2013; Oh-Young & Filler, 2015) The emergence ofdata-driven Artificial Intelligence (AI) as an instrumental force in the educational spheremarks an essential shift in tailoring educational experiences AI’s capabilities align withthe unique learning needs of children with disabilities, promising a new frontier inpersonalised, accessible, and quality education This research explores the innovativeways AI can be harnessed to mitigate the educational disparities that children withdisabilities face, particularly in developing countries.

Building upon this technological advancement, the proposed model in this study seeksto offer a comprehensive, data-driven approach that synergises AI, machine learning,and the specific regional context The model focuses on the non-intrusive collection andanalysis of data, the development of adaptive teaching practices, policy formulation, anda cross-collaborative framework that connects schools, regions, and national entities.The subsequent sections will detail the challenges, scenarios, strategies, and potential

outcomes of integrating AI within the education sector, spotlighting the proposed modelas a beacon for future innovation and inclusivity.

Disability is a condition that limits a person’s ability to perform activities as they wouldlike or what is considered normal (DESA, 2023) This section will initially exploreeducation for disabled children, focusing on developing countries, followed by a fewdeveloped models or solutions According to the WHO (2023), there are many types ofdisability, including physical, sensory, intellectual, and mental health conditions For adetailed view of what a disability is and its models, see UNESCO (2022).

The education of disabled children presents a substantial global challenge.Approximately 1.3 billion people, or 16% of the world’s population, live with disabilities(DESA, 2023) As the largest minority group globally, the growth of this demographic ispropelled by factors such as population growth, medical advancements, and ageingtrends (WHO, 2023) Among those with disabilities, most reside in developing nations,drawing attention to the pervasive issue in lower-income contexts.

The WHO (2022) has highlighted that around 80% of individuals with disabilities live indeveloping countries Women are particularly impacted within OECD countries,reporting higher disability rates in most OECD nations, indicative of substantialsocioeconomic disparities (OECD, 2022) Moreover, the world’s poorest populationsinclude nearly 20% of individuals with disabilities (World Bank, 2018) The intersectionof gender and disability further compounds these disparities, leaving women moresusceptible to marginalisation (Disabled World, 2023) Such inequalities are exemplifiedby elevated risks of abuse, including domestic violence.

2.3 The Status of Children with Disabilities

The situation concerning children with disabilities is gravely alarming, with UNICEF(2021) reporting that around 30% of street youth have some disability, and mortalityrates reaching as high as 80% in certain regions In the context of developing countries,education for disabled children faces profound challenges The WHO (2022) notes thatonly 1 in 5 disabled children have access to education, with UNESCO (2018)emphasising that the education gap is more acute in emerging economies Alarmingly,90% of disabled children in these countries do not attend school, leading to low globalliteracy rates for adults with disabilities and even lower for women with disabilities(UNICEF, 2018).

Segregation, poverty, discrimination, and lack of access to appropriate educationalservices exacerbate these issues, often isolating disabled children from their non-disabled peers The attempts at inclusive education through various approaches, suchas curriculum modifications and assistive technology, have yielded mixed results(Shivashankar & Bakthavatchaalam, 2019) COVID-19 has intensified the struggles ofthis vulnerable group, disproportionately affecting their learning, especially during the

transition to online education (NCPEDP, 2020) The lack of infrastructure, resources,affordability, specialised teaching methods, and societal stigma further hindered thepath to quality education (Peters & Hehir, 2017) Additionally, the research bythe OECD (2022), McLaughlin and Noonan (2017), and Smith et al (2018) delineate themultitude of challenges faced by children with disabilities.

These barriers extend beyond educational hindrances, impeding social interaction andfuture community engagement The lack of suitable accommodations and supportmechanisms translates into lower learning outcomes and stifled skill development.Despite global initiatives towards inclusive education, the pursuit of equitable educationfor disabled children, particularly in developing countries, continues to be a multifacetedand often neglected concern The alarming statistics and the vast disparities in accessto quality education for children with disabilities necessitate a concerted global effort.Policymakers, educational practitioners, and communities must work collaboratively toaddress these inequalities Only through targeted interventions, informed policies, anddedicated resources can the educational rights of all children, regardless of ability, befully realised, thus contributing to a more inclusive and just society.

2.4 AI as a Tool for Inclusion: A Review

Integrating Artificial Intelligence (AI) with wearable technologies revolutioniseseducation, particularly for disabled students in developing countries A few of the AI-based technologies for people with sensory and intellectual disabilities and research onthem are discussed in this section.

Regarding Sensory disabilities, Mukhiddinov & Cho (2021) use smart glasses equippedwith Convolutional Neural Networks (CNN) to provide real-time accessibility for visuallyimpaired students Similarly, Wang et al (2023), in their study, comment on how AIcould be used for obstacle avoidance and assisted reading technology using deeplearning techniques Exploring auditory disability, Lesica et al (2021) comment on howAI technologies in hearing can improve care for common auditory conditions, fosterunderstanding of complex disorders like tinnitus, and develop artificial auditory systems.Using an in-ear onboard training model, Bhowmik et al (2021) developed an on-demand edge AI computing solution that enhances speech clarity Looking at someindustry developments, Kesari (2023) comments on Starkey’s ‘Genesis AI’, which usesa Deep Neural Network (DNN) accelerator engine to analyse and optimise sound in realtime This enables hearing aids to function like the human cerebral cortex, enhancingspeech, reducing noise, and providing a more personalised auditory experience.

Regarding intellectual and mental health conditions, Megat et al (2023) looked at usinga speed-reading tool powered by AI technologies to assist students with ADHD,dyslexia, or short attention spans The tool uses the Multilayer Perceptron algorithm fortext processing and summarisation, using ‘Bionic Reading’ principles toenhance readability For children with autism, Haber et al (2020) use a wearable deviceas a social interaction aid This focuses on addressing deficits in eye contact, facialexpression recognition, and social interaction Similarly, a randomised controlled trial by

Fahimi et al (2019) found that the EEG headband with CNN effectively reducedinattention and hyperactivity symptoms of children and adolescents with ADHD Crippaet al (2015) explored using the ‘Support Vector Machine’ to identify children with AutismSpectrum Disorder by analysing upper-limb movements such as reach, grasp, and drop.AI’s potential to personalise content delivery is unlocking new educational avenuestranslating and adapting content to local customs, languages, and histories while beingaware of possible biases (Kazimzade et al., 2019) Virtual classrooms empowered by AIare breaking linguistic barriers, and AI-driven administrative automation frees educatorsto focus on teaching, fostering continuous improvement through a self-determinationtheory design approach (Xia et al., 2022) AI aids in inclusive education, offering tailor-made content formats and real-time adaptation, with emerging technologies monitoringstudents’ emotional states to enhance learning (Balachandran & Rabbiraj, 2023).

Most research in this field has primarily occurred within developed countries, where thesocio-economic landscape differs significantly However, there is a noticeable scarcityof studies addressing developing and implementing management practices and policiesfor children with disabilities through a data-driven model within evolving economiccontexts, particularly in developing nations This research aims to bridge this critical gapby focusing primarily on developing countries The study endeavours to design acomprehensive model and thoroughly investigate the numerous prospects andobstacles associated with its implementation within developing country contexts Bydoing so, it aspires to contribute significantly to formulating effective managementpractices and policies tailored to the unique needs and challenges of children withdisabilities in these regions.

2.5 Proposed Model

This section focuses on the newly conceived three-tiered data-driven model underdevelopment, explicitly aiming to enhance teaching and learning for children withdisabilities within the framework of emerging economies This model is centred aroundcreating a system that integrates management and policy, providing a comprehensiveapproach to education for those with special needs (Figure 2.1).

Figure 2.1: Proposed model of the three-tiered data-driven model.

The proposed data-driven model, as depicted in Figure 2.1, is segmented into threesections The initial phase involves the non-intrusive collection of data at the schoollevel from students with disabilities, aiming to pinpoint specific needs based ondemography and individual requirements This data is relayed to regional orgovernmental bodies in conjunction with similar information gathered from otherinstitutions The broad spectrum of data thus accumulated enables a thorough analysis,facilitating the creation of policies and regulations tailored to the region’s uniquecharacteristics.

Furthermore, the model accommodates data from various regions or states, creating acentralised repository This aggregation of diverse data concerning teachingmethodologies, their effectiveness, and outcomes fosters a detailed examination,culminating in the formulating of informed practices and policies The interplay betweennational and regional data contributes to a coherent strategy where national insightsinform regional and school-level policies In contrast, regional policies reciprocate byshaping data-driven practices at the institutional level.

The embodiment of this model not only empowers individual schools and regions tocultivate their distinct practices and policies but also ensures that strategies arerecorded and disseminated across the regional and national levels This collaborativeapproach, underpinned by the proposed model, is poised to shift the education ofchildren with disabilities, aligning practices and policies with the specific needs andcontexts of the children they are designed to support The following section will look atthe model in detail, exploring the various non-intrusive data collection, analysis of thedata, the training of the algorithm and how it incorporates feedback mechanisms to

iterate the data-driven process within the three-tier model for creating effectiveeducation management policies and practices.

2.6 Research Methodology

The research methodology is divided into two main phases: the data collection, whichintegrates quantitative and qualitative insights, and a comprehensive data analysisphase employing advanced machine learning algorithms.

During the initial phase, data are meticulously collected at individual, regional, andnational levels The method leverages wearables to acquire real-time physiological dataand standardised assessments to monitor academic and social progressions at theindividual or school level Furthermore, qualitative feedback is sought from keystakeholders, such as students, parents, and teachers, furnishing a well-rounded viewof the educational process Moving to a broader scope, regional-level data analysisamalgamates information from various schools to discern regional trends, which wouldbe vital in crafting policies that are sensitive to local needs yet have the potential to bescaled nationally Lastly, a macro-analysis is conducted at the national level, whichharnesses the accumulated regional data to foster a unified and data-informededucational strategy.

The second phase delves deep into data analysis and policy formulation Here, machinelearning models, including Decision Trees, Random Forests, and Transformer-basedmodels, are utilised to dissect and comprehend intricate data sets, providing a versatileanalytical framework Subsequently, the policy-making process is facilitated throughdata-driven dashboards at different administrative levels These dashboards assist inshaping teaching methodologies at the local level and formulate grounded policyrecommendations at the regional level based on a thorough analysis of quantitative andqualitative feedback At the national spectrum, aggregating and synthesising a vast poolof textual data assist policymakers in identifying broader trends, which in turn helpshape nationwide strategies and reforms.

In addition to the methodological aspects, this research accentuates the necessity foropen science and adhering to stringent ethical guidelines, promoting transparency andreproducibility in research This approach envisages fostering multi-disciplinarycollaborations for further refinement and longitudinal studies Moreover, a tieredgovernance structure is proposed, incorporating local educational bodies, regionaldepartments, and national entities to warrant a coherent and effective strategyimplementation backed by multidisciplinary inputs.

2.7 Data Collection

This section discusses the data collection at all the levels based on the model At theindividual or school level, three types of data will be collected:

1 Quantitative Data from Sensors: This data, acquired through wearables like heartrate monitors and temperature sensors, will offer insights into thephysiological and emotional states of students, enabling real-time adjustments tolessons and interventions.

2 Quantitative Scores from Standard Assessments: These scores and metrics willbe derived from standardised assessments designed to address each child’sunique social and academic needs.

3 Qualitative Feedback: Qualitative feedback will be gathered from children,teachers, parents, and caregivers This feedback mechanism will complementthe quantitative data, ensuring active stakeholder engagement in the educationaljourney.

The initial phase of our study involves the continuous collection of data from studentswith disabilities in the classroom The data can be collected in schools or at home,depending on the child’s requirements Based on the model, data will be collected usinga variety of sensors, including heart rate monitors, temperature sensors, and otherphysiological sensors At the same time, students are engaged in the learning process.Non-intrusive wearables designed for this research or easily available devices likesmartwatches can be used.

In addition to sensor data, quantitative data based on a standardised set ofassessments tailored to the unique learning requirements of individual children will becollected This data will include scores and metrics from standard assessmentsconsidering the children’s unique social and academic needs These assessments willencompass functional capabilities, social interactions, and academic achievements.Combining these standardised tests with continuous monitoring through wearables willprovide a comprehensive understanding of students’ learning abilities, a critical factor indesigning personalised interventions.

In addition to the quantitative data, more holistic qualitative feedback will be gatheredfrom children (if appropriate), teachers, parents, and caregivers as required Thequalitative data ensures that every stakeholder remains engaged and activelyparticipates in the educational journey Qualitative data offers rich, contextual insightsthat lend depth to the statistical findings, helping researchers understand the ‘why’behind the numbers For example, while test scores (quantitative data) may indicate adecline in performance, interviews with children and teachers (qualitative data) mayreveal underlying reasons, such as a specific technological requirement or a lack ofsufficient supporting mechanisms.

This qualitative data can be further used to determine the teaching and learningpractice’s effectiveness, make necessary changes to the curriculum iteratively, andinfluence the policy on various levels within the three-tiered proposed model Qualitativefeedback from these groups is essential for understanding the lived realities of the

educational experience beyond what can be captured through quantitative metricsalone The three data types will be collected at the individual level within the schools.Similarly, data will be collected from various participating schools at the regional level,which will then be aggregated to form a regional data repository Analysing trends atregional levels makes it possible to design interventions sensitive to local needswhile also scaling effective strategies to a national level This fosters a dynamic,responsive policy environment that is both locally relevant and nationally effective.Regional and national policy evaluation is a scale-up of insights, moving from granular,school-level analyses to broader, systemic evaluations Data repositories that aggregateinformation from various regional schools allow researchers and policymakers to identifycommon trends, challenges, and successes that may be specific to that geographicalarea.

Once regional data repositories have been analysed, they serve as an empirical basisfor designing and implementing region-specific interventions These interventions canthen be piloted on a smaller scale before broader implementation Such a targetedapproach ensures that policies are finely tuned to local needs, sociocultural contexts,and existing infrastructures, thereby enhancing their likelihood of success.

While regional repositories offer insights into local trends, aggregating theserepositories at a national level facilitates the formation of overarching educationalstrategies A national data repository offers a macroscopic view that captures theeducational landscape, including regional disparities or consistencies National-levelanalyses can help policymakers to identify commonalities and differences, resourceallocation and policy harmonisation The suggested approach involves identifyingcommon challenges and achievements at a national level while recognising regionalvariations, which enables the development of a more nuanced policy framework This,in turn, facilitates informed resource allocation decisions, such as deploying teachers,educational materials, or specialised training programs where they are most needed.Additionally, the proposal advocates establishing national benchmarks or standardsgrounded in empirical evidence, which can serve as a basis for evaluating regional orlocal educational performance Lastly, it emphasises the importance of integrating andharmonising educational policies to achieve a degree of standardisation across regionswhile accommodating local adaptations.

Incorporating wearable sensors in educational settings offers a granular approach topersonalised learning, particularly for students with disabilities These sensors collectreal-time physiological and emotional data, enabling tailored educational experiences.This data is then seamlessly integrated into a mobile-based application connected tonearby edge devices, which aggregate the information from multiple students andschools in the region Edge computing enhances data processing efficiency andenables quicker decision-making by reducing latency and bandwidth usage, thusmaking it feasible for real-time interventions or policy adjustments at local, regional, oreven national levels However, the methodology faces challenges, including sensoraccuracy, high data volume, untrained usage and ethical considerations like informed

consent and prioritising the privacy and anonymity of individuals by subjecting them tomasking and homomorphic encryption Interdisciplinary collaboration and furtherlongitudinal studies could significantly refine this data collection technique.

2.8 Analysis and Policy Making

Quantitative sensor data are used as the features Features are training example datapoints such as the data from many sensors used to infer an individual’s emotional andphysiological state The data from the standardised assessment scores are used as theground truth label With this data in hand, we aim to create a comprehensive dataset.The data collected from the sensors will be the independent variables, whereas thescores from the standardised assessments are the dependent variable quantitatively.The data can be analysed using various algorithms such as ‘Decision Trees’, ‘RandomForests’, and ‘Transformer’-based models Given the complex nature of the data andthe tasks, advanced ML models should be considered to enhance the analyticalframework A few algorithms that can be used for analysing the data at different levelsinclude ensemble learning methods such as ‘Gradient Boosting Machines’ and‘XGBoost’, especially for cases with missing data – suited to the current requirement.For analysing the quantitative data, ‘Recurrent Neural Networks’ and ‘Long Short-TermMemory Networks’ can offer insights into trends over time in a student’s physiologicaldata In addition, more specific variants like BERT (Bidirectional et al fromTransformers) or GPT (Generative Pre-trained Transformer) can be employed forcomplex tasks, such as understanding the context in which physiological changes occurin the students based on textual data (e.g., classroom transcripts or notes) These MLmodels can be selectively deployed depending on the specific data characteristics andthe analytical requirements, thereby adding significant robustness and versatility to theanalytical framework Furthermore, considering security and data privacy, we could alsoemploy federated learning models to train on decentralised data across schools whilekeeping raw data localised This aligns well with the noted need for localised decision-making in educational interventions.

Qualitative feedback from the school level can be integrated into the dashboard toprovide context to the quantitative KPIs For instance, teacher observations aboutstudent behaviour can be juxtaposed with physiological data from wearables to offer aricher understanding of student well-being Natural Language Processing (NLP)algorithms such as ‘Topic Modeling’ or ‘Sentiment Analysis’ can summarise thisqualitative feedback, allowing immediate and actionable insights The dashboard couldflag quantitative and qualitative data discrepancies, prompting further investigation.Overall, a web or mobile app dashboard at the local school level will display KeyPerformance Indicators such as student engagement, stress levels and attendancepatterns A set of metrics around academic performance collected through the appcould help teachers tailor their teaching methods.

For regional educational authorities, an advanced dashboard version could aggregatedata across multiple schools KPIs like average student well-being, teacher

effectiveness, and resource utilisation could be monitored Advanced models likeGradient Boosting Machines or Bayesian Models could analyse regional trends, offeringpolicy recommendations such as resource allocation or curricular changes.Regionally, qualitative data can provide insights into varying educational methods,resource utilisation, and school stakeholder satisfaction ML techniques like ‘LatentDirichlet Allocation’ can cluster similar feedback, helping educational authoritiesunderstand common themes or issues in different schools This will enable them to craftpolicies for the region-specific challenges.

On a national scale, the dashboard could incorporate broader metrics like literacy rates,regional comparisons, and national test scores Transformer-based models couldsynthesise large-scale textual data, such as feedback from educators nationwide, toidentify overarching trends and challenges This would enable policymakers to makedata-driven decisions on a grander scale, such as nationwide curricular reforms ortargeted teacher training programs On a national scale, the synthesis of qualitativefeedback can inform broader educational strategies Advanced NLP models, such asTransformer-based algorithms, can process large volumes of text data, identifyingoverarching trends and needs that may not be apparent through quantitative metricsalone For instance, widespread concerns about exam stress could lead to policyrecommendations about exam reform or mental health support in schools.

In principle, all the non-sensitive data and resource materials would be made publiclyavailable These include the research designs, draft ethical forms, codes used,schematics and experiments set up and so on for further scrutiny and researchpurposes The final AI models, results and policies will also be opened up forreproducibility and adaptation for the community and similar developing countries.

At the local or school level, educational institutions are responsible for the initial datacollection, monitored by local education authorities and in consultation with parents andcaregivers Local bodies will thus handle the daily data acquisition, ensuring it alignswith ethical and procedural norms Moving up to the regional or state level, the StateEducation Department takes the helm They are charged with aggregating the collecteddata into regional repositories and are accountable for its oversight Regional healthcareinstitutions and non-governmental organisations (NGOs) play supplementary roles,aiding data interpretation and liaising between the government and the community TheMinistry of Education oversees the entire framework at the national level, includingmaintaining a national data repository and formulating educational benchmarks andstandards National Data Protection Authorities ensure compliance with data privacyregulations Ministries of Health, advisory committees, and academic researchinstitutions provide multidisciplinary expertise to inform and evaluate the implementedpolicies Together, these entities contribute to a coherent and effective governancestructure.

2.9 Barriers to Implementing this Model in Developing Countries

This data-driven model can create a significant impact in the lives of disabled childrenby offering innovative solutions tailored to their unique needs and challenges However,implementing this in developing countries is fraught with numerous barriers In thedeveloping countries, a few of the often-cited barriers include the lack of technology,resources and infrastructure, need for qualified teachers, social stigma (Kuper et al.,2018; Oh-Young & Filler, 2015) The situation is further complicated in countries likeIndia, where the diversity of regional languages presents additional hurdles Moreover,economic constraints and the lack of established ethical principles and legislationconcerning AI (Chatterjee, 2020), particularly in relation to disabled children, exacerbatethese difficulties.

This section explores the various themes of barriers: Technological, cultural, economic,administrative and regulatory barriers Along with identifying the themes, this sectionalso addresses how the interplay of these themes influences the adoption and effectiveuse of the data-driven model.

A prominent obstacle is limited access to technology (Shivashankar &Bakthavatchaalam, 2021), particularly in rural areas of developing countries, whereessential infrastructure and equipment for AI-based solutions are often lacking.Moreover, WHO and UNICEF (2022) emphasise the global disparity in the availability ofassistive technology in low and middle-income nations Dias (2015) further highlightstechnical issues, security concerns (Yathiraju et al., 2022), and a digital divide that limitsaccess to smart devices for children and educators Schools and governments mustensure equitable access to these devices to mitigate this Given the increasingaffordability of mobile phones in developing countries (IFC, 2020; Silver et al., 2019),the model needed to be developed with easy access through mobile phones, thusincreasing accessibility.

Along with this, the lack of access to special schools is another issue Schools indeveloping countries are often not accessible to children with disabilities This may bedue to physical barriers, such as lack of wheelchair ramps or accessible toilets, or policybarriers, such as lack of inclusive education policies (Dyssegaard & Larsen, 2013; Oh-Young & Filler, 2015) In addition, the data collected needs to be harmonised (Mont etal., 2022) for the model both at the statistical analysis system and the administrativesystem., but this could pose a challenge at a regional and national level Even thoughthis has been considered in devising this model, it must be carefully implemented.

Digital literacy and expertise gaps, as noted by IFC (2020) and Bapna et al (2020), aremajor barriers to implementing this model or similar AI-based solutions in developingcountries The lack of qualified teachers capable of using technology with disabledchildren is a common theme that is reported by several works, including WHO andUNICEF (2022), IFC (2020), Shalini and Shipra (2020) and Hashemi et al (2020) Theshortage of trained teachers, primarily those proficient in educating childrenwith disabilities and those who can correctly use the wearables and collect proper data,

threatens to deprive these children of the support they require for academic and emotional development (Hayes & Bulat, 2017; Kuper et al., 2018; Szumski et al., 2017).So, implementing this model and making a success of it requires investigation This isnoted to be one of the significant barriers, so special needs teaching and theGovernments must look at the funding requirements There is a need to provide moretraining for teachers on how to work with children with disabilities This training shouldcover both the academic and the social-emotional aspects of education.

socio-The cost associated with special education is comparatively higher than that of regularschool, making it difficult for middle-class and low-income families to afford (Shalini &Shipra, 2020; Lakshminarayana et al., 2019; Senjam et al., 2021) Many families indeveloping countries cannot afford to send their children to school, even if the school isfree This is especially true for families with children with disabilities, as they may incuradditional costs for special education or assistive devices (UNICEF, 2018; WHO,2011; World Bank, 2018) Governments need to invest more in education, especially forchildren with disabilities This includes providing financial assistance to families, makingschools more accessible, and training more teachers (UN, 2019; WHO, 2011).

Additionally, a pervasive social stigma is associated with disability and special schoolsin developing countries, constituting a substantial barrier (Kuper et al., 2018; Oh-Young& Filler, 2015) Negative attitudes towards disability can lead to discrimination againstdisabled children within educational institutions and communities This can make itdifficult for children with disabilities to feel welcome and included in school (Kuper et al.,2018; Oh-Young & Filler, 2015) There is a need to change attitudes towards disabilityin developing countries This is an essential barrier in implementing this model, as thedata collected might need to be revised To overcome this, concerted efforts are neededto change societal perceptions through awareness campaigns, community leadertraining, and media initiatives with appropriate cultural sensitivity.

There are also barriers to implementing this model using ML Even though ML offerssignificant potential for extracting valuable insights from continuous physiologicalmonitoring in healthcare, as noted by Rush et al (2019) However, this potential iscounterbalanced by regulatory and medico-legal issues, necessitating rigorousevaluation and clear demonstration of value Simultaneously, the reliance on potentiallyunreliable datasets can lead to machine bias stemming from errors or incomplete datainput The operationalisation of ML models across various levels in the model alsoposes a challenge, as it demands substantial capital investment and the presence ofadequately trained staff.

Moreover, the often-opaque nature of ML algorithms – the often quoted ‘black box’factor – makes understanding the underlying rationale for specific conclusions difficult.The introduction of accurate learning ML models further compounds these issues,raising complex questions about decision-making processes and potential adverseoutcomes In this model, not all the measures used are quantitative, with some of thembeing qualitative, especially the feedback from the children, teachers, parents, schoolmanagement and the other stakeholders in terms of their learning and functional skills,

the subjective nature of such data could impact the results, all these would have to becarefully considered while training the model.

Finally, the absence of relevant legislation, particularly in AI and data usage withvulnerable populations, represents a critical barrier (Jobin et al., 2019) Governmentsmust adopt inclusive education policies, meaning all children can attend the sameschools regardless of their abilities This will help to break down barriers and create amore inclusive environment for all children (Dyssegaard & Larsen, 2013; Oh-Young &Filler, 2015).

Concluding this section, rather than identifying the barriers in isolation, it is essential tocomprehend how they collectively impact individuals with disabilities and their families.This holistic understanding becomes essential for devising and executing efficientinterventions to enhance access to primary healthcare Addressing these barriersrequires a multidimensional approach involving governments, internationalorganisations, academia, and the private sector to create inclusive and accessible AIsolutions To provide holistic support for the children, changes need to happen not onlyin the teaching practice but also in the curriculum design, textbook and online material,teacher education, school facilities, and community inclusion, thus helping formappropriate management practices.

2.10 Future Works

This study encompasses a convergence of four key domains: data-driven computingand AI, experts in the field of disabilities, educators, and policymakers While it might beeasy for future researchers to narrow their focus on the AI aspects of this research, theymust approach the children with empathy and adopt a sensitive approach whilstgathering data and formulating policies grounded in the data Furthermore, given theirvulnerable nature, the chapter advocates the presence of either a parent or a trainedexpert during interactions with children, given their vulnerable nature This work aims tocreate a data-driven model for emerging economies where social dynamics differ fromthose in developed or underdeveloped countries The framework should not begeneralised in emerging economies, as political, geographical, and cultural dimensionsvary within regions and countries As a result, the framework should consider thesedifferences and produce context-specific practices derived from the data.

Future research could consider validating the general framework using quantitative andqualitative data in longitudinal studies This work focuses on wearables, learningassessment, and a particular type of data analysis However, AI’s swiftly evolvinglandscape necessitates considering novel data collection and analysis methodologiesand the results tested with other AI methods Future research should also consider theimplications of data collection from disabled children on their long-term development.Additionally, this research should have considered the political will among emergingeconomies for comprehensive regional and national research initiatives Future workshould investigate this in detail.

In conclusion, data-driven decision-making can accelerate inclusive education forpeople with disabilities By collecting and analysing data on student outcomes,educators can identify areas where improvements are needed and develop targetedinterventions The framework developed can help ensure that all students, regardless oftheir abilities, have the opportunity to succeed in school At the same time, the chapterhas also identified various challenges in implementing such a system in developingcountries Despite these challenges, data-driven decision-making has the potential tomake a significant difference in the lives of students with disabilities By overcoming orreducing the challenges, educators and policymakers can use data to improve studentoutcomes and create more inclusive schools for all and overall in achieving SDG 4.References

Antoninis, M., April, D., Barakat, B., Bella, N., D’Addio, A C., Eck, M., Endrizzi, F.,Joshi, P., Kubacka, K., McWilliam, A., Murakami, Y., Smith, W., Stipanovic, L., Vidarte,R & Zekrya, L (2020) ‘All means all: An introduction to the 2020 Global Education

Monitoring Report on inclusion’, PROSPECTS, 49(3), pp 103–109. a, b

Balachandran, M., & Rabbiraj, C (2023) Artificial intelligence: a support system ininclusive education International Journal of Knowledge and Learning. →

Bhowmik, A K., Fabry, D A., Armour, P., Berghel, H., Charette, R N., & King, J L.(2021) Hear, now, and in the future: Transforming hearing aids into multipurposedevices Computer, 54(11), 108–120 https://doi.org/10.1109/MC.2021.3105151 →Chatterjee, S (2020) “AI strategy of India: policy framework, adoption challenges andactions for government”, Transforming Government: People, Process and Policy, Vol.14 No 5, pp 757–775 https://doi.org/10.1108/TG-05-2019-0031 

Crippa A, Salvatore C, Perego P, Forti S, Nobile M, Molteni M, et al Use of machine

learning to identify children with autism and their motor abnormalities J Autism Dev

Disabled World (2023) Disability Statistics Retrieved from world.com/disability/statistics/ →

https://www.disabled-Dyssegaard, A., & Larsen, K (2013) Inclusive education: A review of the evidence.Copenhagen: Danish Institute for International Studies. a, b, c, d

Fahimi, F., Zhang, Z., Goh, W.B., Lee, T.S., Ang, K.K and Guan, C (2019) subject transfer learning with an end-to-end deep convolutional neural network for EEG-based BCI Journal of neural engineering, 16(2), p 026007. 

Inter-Haber, N., Voss, C., Daniels, J., Washington, P., Fazel, A., Kline, A., De, T., Winograd,T., Feinstein, C., & Wall, D P (2020) A Wearable Social Interaction Aid for Childrenwith Autism Retrieved from http://arxiv.org/abs/2004.14281v1 →

Hashemi, G., Wickenden, M., Bright, T., & Kuper, H (2020) Barriers to accessingprimary healthcare services for people with disabilities in low and middle-incomecountries: A meta-synthesis of qualitative studies Disability and Rehabilitation, 44(8),1207–1220 doi: 10.1080/09638288.2020.1817984 →

Hayes, D., & Bulat, M (2017) The impact of inclusive education on the academicachievement of students with disabilities: A systematic review Review of Education35(4), 375–407. a, b

International Finance Corporation: WEF (2020) Artificial intelligence in emergingmarkets: Opportunities, trends, and business models Geneva: World EconomicForum. a, b, c

Jobin, A., Ienca, M & Vayena, E (2019) The global landscape of AI ethics guidelines.Nat Mach Intell 1, 389–399 https://doi.org/10.1038/s42256-019-0088-2 

Kazimzade, G., Patzer, Y., & Pinkwart, N (2019) Artificial Intelligence in EducationMeets Inclusive Educational Technology – The Technical State-of-the-Art and PossibleDirections. →

Kesari, G (2023) ‘AI vs Hearing Loss: The Battle That is Transforming Millions ofLives’, Forbes Available at:https://www.forbes.com/sites/ganeskesari/2023/07/28/ai-vs-hearing-loss-the-battle-thats-transforming-millions-of-lives/ →

Kuper, H., et al (2018) The impact of inclusive education on the social-emotional being of students with disabilities: A systematic review Review of Education 36(1), 3–30. a, b, c, d, e

well-Lesica, N.A., Mehta, N., Manjaly, J.G et al Harnessing the power of artificialintelligence to transform hearing healthcare and research Nat Mach Intell 3, 840–849

McLaughlin, M E., & Noonan, J R (2017) The role of teachers in inclusive education.Review of Educational Research, 87(1), 1–32. →

Megat, K.I.Z, Hafizzul, A.N.B.Y.M and Bin, B.A.M.M.M (2023) Speed Reading ToolPowered by Artificial Intelligence for Students with ADHD, Dyslexia, or Short AttentionSpan arXiv preprint arXiv:2307.14544. 

Mohammad, Sheikh Sajid, & Huzina Saheal (2023) “A Sneak Peek into the Future ofArtificial Intelligence in Education: Opportunities and Challenges.” In DigitalTransformation in Education: Emerging Markets and Opportunities, edited by C.Krishnan, m Babbar and Singh G, pp 207–222 Singapore: Bentham. 

Mont, Daniel, Jennifer Madans, Julie D Weeks, & Heidi Ullmann (2022) “Commentarydiscusses harmonising disability data to improve disability research and policy.”Disability 41 (10): 1442–1448 https://doi.org/10.1377/hlthaff.2022.00479 →

Mukhiddinov, M & Cho, J (2021) Smart Glass System Using Deep Learning for the

10(22):2756 doi.org/10.3390/electronics10222756 

NCPEDP (2020) Locked down and Left behind: A Report on the Status of Personswith Disabilities in India during the COVID-19 Crisis New Delhi, India: NCPEDP. →OECD (2022), Disability, Work and Inclusion: Mainstreaming in All Policies andPractices, OECD Publishing, Paris, https://doi.org/10.1787/1eaa5e9c-en a, b

Oh-Young, I., & Filler, J (2015) Inclusive education in low- and middle-incomecountries: A systematic review of the evidence International Journal of Disability,Development and Education 62(4), 348–367. a, b, c, d, e, f

Peters, S., & Hehir, T (2017) Barriers to inclusive education: A review of the literature.Disability & Society, 32(2), 171–185. →

Rush, B., Celi, L.A & Stone, D.J Applying machine learning to continuously monitoredphysiological data J Clin Monit Comput 33, 887–893 (2019). →

Senjam, S S., Foster, A., & Bascaran, C (2021) Barriers to using assistive technologyamong students with visual disability in schools for blind students in Delhi, India.Disability and Rehabilitation: Assistive Technology, 16(7), 802–806. →

Shivashankar, K and Bakthavatchaalam, V., (2019) “Enhancing the learning process ofpeople with special needs using wearables and Artificial Intelligence.” Mobile LearningWeek, Paris. →

Shivashankar, K and Bakthavatchaalam, V., (2021) Smart Municipal GovernanceSystem Using Confluence of Technologies in a Developing Country Scenario InProceedings of the 14th International Conference on Theory and Practice of ElectronicGovernance (pp 535–537) doi: 10.1145/3494193.3494277 →

Silver, L., Aaron S., C J., Kyle T., J., J., Monica A & Lee R (2019) “MobileConnectivity in Emerging Economies.” Pew Research Center, March7 https://www.pewresearch.org/internet/2019/03/07/mobile-connectivity-in-emerging-economies/ →

Smith, C J., Turnbull, A M., & Robertson, A C (2018) Parental perspectives oninclusive education: A review of the literature Int Journal of Inclusive Education, 22(8),883–901. →

Smith, J., Johnson, K., & Lee, T (2020) Artificial Intelligence in Inclusive Education: AComprehensive Review Journal of Educational Technology, 12(3), 45–60. 

Szumski, G., et al (2017) Inclusive education: A systematic review of the literature.European Journal of Special Needs Education 32(1), 4–25. a, b

The European Agency for Special Needs and Inclusive Education (2018) Inclusiveeducation: A systematic review of the evidence Odense: European Agency for SpecialNeeds and Inclusive Education. 

UN 2019 UN Disability and Development Report – Realizing the SDGs by, for and withpersons with disabilities UN (NY). →

UNESCO, (2022) Welcoming Diversity in a Learning Environment: Teacher’s handbookfor inclusive education, Paris. →

UNESCO (2017) Education for Sustainable Development Goals: Learning objectives.UNESCO (Paris) https://www.voced.edu.au/content/ngv:77653. →

UNESCO Institute for Statistics (UIS) (2018) Education and disability: analysis of datafrom 49 countries Information Paper No 49. 

UNHCR (2014) The convention on the rights of persons with disabilities New York.UN. →

UNICEF (2017) The state of the world’s children 2017: Children in a digital world NewYork: UNICEF. 

UNICEF (2018) The state of the world’s children 2018: Children in a digital world NewYork: UNICEF. a, b

UNICEF (2021) Policy guidance on AI for children (NY: UNICEF). →

Wang, J., Wang, S & Zhang, Y., 2023 Artificial intelligence for the visually impaired.

at: https://doi.org/10.1016/j.displa.2023.102391 →

Wang, N., Tonko, P., Ragav, N., Chungyoun, M., & Plucker, J (2022) A Perspective onK-12 AI Education Retrieved from https://doi.org/10.21300/23.1.2023.2 

WHO (2022) Global report on health equity for persons with disabilities Geneva: WorldHealth Organization. a, b

WHO (2023) Addressing health inequities faced by persons with disabilities to advanceuniversal health coverage Technical brief Geneva: World Health Organization. →

health https://www.who.int/en/news-room/fact-sheets/detail/disability-and-health →WHO 2011 World report on disability Edited by: WHO and World Bank Geneva. a, bWorld Bank (2018) Disability and development: The World Bank’s approach to disabilityin the 2030 development agenda Washington, DC: World Bank. a, b

Xia, Q., Chiu, T K F., Lee, M K., Sanusi, I T., Dai, Y., & Chai, C (2022) A determination theory (SDT) design approach for inclusive and diverse artificialintelligence (AI) education Computers & Education. →

self-Yathiraju, N., Jakka, G., Parisa, S K., & Oni, O (2022) Cybersecurity Capabilities inDeveloping Nations and Its Impact on Global Security: A Survey of Social EngineeringAttacks and Steps for Mitigation of These Attacks In M Dawson, O Tabona, & T.Maupong (Eds.), Cybersecurity Capabilities in Developing Nations and Its Impact onGlobal Security (pp 110–132) IGI Global https://doi.org/10.4018/978-1-7998-8693-8.ch007 →

Chapter 3 The Role of Artificial Intelligence in the Emerging Digital Economy Era

Wasswa Shafik

Artificial Intelligence (AI) is changing the current digital economic era (DEE), changinghow the public lives, works and interacts with technology AI has various types,including rule-based machine learning (ML) and neural network AI AI can improveefficiency, productivity, and decision-making but poses challenges, including jobdisplacement and workforce transformation, bias and ethics concerns, and data privacyand security This chapter presents a comprehensive overview of AI in this DEE,demonstrating the significance of AI in the DEE Types of AI are demonstrated,including rule-based AI, ML, AI, and Neural Networks (NNs) AI applications in the DEEare surveyed, for instance, chatbots and virtual assistants, image and speechrecognition, predictive analytics, autonomous vehicles, and intelligent healthcare Thefuture of AI in the DEE is promising, with advancements in AI technology, increasedadoption, and potential impacts on society and the economy The study further presentsthe benefits of AI in the DEE, followed by the challenges that AI encounters as per the

current technological standard It is crucial to consider AI’s implications and guaranteethat its development and deployment are ethical, secure, and beneficial for everyone.Finally, the future of AI in this emerging era is depicted As AI continues to evolve, it willundoubtedly bring benefits and challenges, and it is essential to develop AI responsiblythat aligns with ethical values and protects individual rights.

Keywords: Decent Work and Economic Growth, Artificial Intelligence, Digital Era, Machine

Learning, Privacy and Security, Ethics Concerns, Responsible Consumption andProduction, Industry, Innovation, and Infrastructure,

Artificial Intelligence (AI) has developed as a transformative technology in the emergingdigital economic era (DEE), impacting various aspects of our lives, including work,entertainment, and communication The advancements in AI have opened up newpossibilities for automation, data analysis, and decision-making, leading to increasedefficiency and productivity in various industries (Chang, 2022) Incorporating AI inthe digital age presents difficulties, including employment displacement, bias, moralissues, data protection, and security Rule-based AI, machine learning (ML) AI, and NNAI are diverse technologies with specific uses in various fields (Bartosik-Purgat &Filimon, 2022).

AI is becoming increasingly common in contemporary culture, with several applicationsin industries like customer service, banking, healthcare, and transportation Thepotential of AI in the digital age to automate tasks and increase productivity andefficiency is one of its significant advantages AI, for instance, can automate routinejobs, freeing staff members to concentrate on more complicated duties that call forimagination and critical thinking (Fortuna et al., 2020) AI can also analyze enormousvolumes of data and spot patterns humans might have missed, enhancing forecastingand decision-making.

Even though AI has many advantages, it also presents several problems in the digitalage Workforce transformation and job displacement are two of the biggest problems.Some operations may disappear due to AI automation, so the workforce should adapt tonew positions Concerns about prejudice and ethics are also present since AIalgorithms may reflect the biases of their designers and produce unfair results (Nugrahaet al., 2022) AI poses data security and privacy issues because businesses mustensure that client data is safeguarded and not misused (Yujie et al., 2022) Withimprovements in AI technology, greater acceptance, and possible effects on society andthe economy, the future of AI in the DEE looks bright However, it is crucial to approachAI development responsibly, giving individual rights and ethical considerations a toppriority.

Government agencies, business operatives, and academic institutions are all involved inseveral efforts in using AI Governments, for instance, fund AI research anddevelopment to foster innovation and economic development while simultaneouslycreating rules and policies to guarantee that AI development aligns with moral principles

and protects individual rights Industry participants are also creating ethical frameworksand standards for the creation and application of AI, encouraging openness andresponsibility Academic institutions are investigating the moral implications of AI andcreating curricula and training programs for anyone who wants to work in the field(Abutabenjeh et al., 2022) Recognizing AI’s possible hazards and difficulties, such asjob displacement, bias, and privacy problems, is crucial.

3.1.1 The Contribution of the Chapter

This chapter provides the following contributions:

 The chapter provides insight into the evolving landscape of AI in the digital ageand its impact on individuals, enterprises, and society.

 Presents different types of AI, including rule-based AI, ML AI, and neural networkAI.

 Illustrates the applications of AI in digital generation, including chatbots andvirtual assistants, image and speech recognition, predictive analytics,autonomous vehicles, and healthcare.

 Demonstrates the benefits of AI in the emerging DEE, including increasedefficiency and productivity, improved decision-making, enhanced customerexperience, and cost reduction.

 Finally, the study provides some DEE-selected merits and illustrates policyimplications and recommendations.

3.1.2 Chapter Organization

The rest of this study contains the following sections: Section 1.2 entails different typesof AI, including rule-based AI, ML AI, and NN AI Section 1.3 presents the applicationsof AI in digital generation, including chatbots and virtual assistants, image and speechrecognition, predictive analytics, autonomous vehicles, and healthcare, among others.Section 1.4 examines and presents the benefits of AI in the DEE, including increasedefficiency and productivity, improved decision-making, enhanced customer experience,cost reduction, and others Finally, Section 1.6 presents the recommendations and thechapter conclusion.

3.2 Artificial Intelligence Types

AI can be broadly categorized into three main types: rule-based AI, machine learning AI,and NN AI, demonstrating the application and uses as presented in this section.

3.2.1 Rule-Based Artificial Intelligence

Rule-based AI is also known as expert system and knowledge-based AI These AIsystems are intended to adhere to and make decisions based on predefined rules.Rule-based AI is advantageous when the decision-making process is well-defined andcan be programmed within the system Rule-based AI includes chatbots, virtualassistants, and recommendation systems as examples (Li et al., 2020) Siri is a well-known example of a rule-based AI that can answer queries, make recommendations,and perform tasks based on predetermined rules Based on predefined rules, Amazon’svirtual assistant Alexa can perform various duties, including setting alarms, playingmusic, and placing online orders (Andreadis et al., 2022).

3.2.2 Machine Learning and Artificial Intelligence

ML AI is designed to learn and enhance performance based on experience, and thealgorithms are trained on massive data The system can make decisions based on thepatterns and insights it has learned from the data ML-AI is useful when the decision-making process is complex and complex to program explicitly (Kollmann et al., 2022).Facial recognition technology is an example of ML and AI for security and lawenforcement purposes The system is trained on large face datasets and can effectivelyidentify individuals based on facial characteristics.

3.2.3 Neural Networks and Artificial Intelligence

Neural network AI is a ML form that simulates the human brain’s structure andfunctionality AI-assisted NNs is designed to identify patterns and relationships in databy establishing links between various data elements (Trisiana et al., 2019; Jun et al.,2021) This form of AI is appropriate when the data is complex and human interpretationis difficult The image recognition technology used in self-driving vehicles is an exampleof NN artificial intelligence The system is educated on enormous quantities of imagesof roads, traffic signals, and other objects to identify and navigate the environment(Boonmoh et al., 2021) Another example is the recommendation system used byNetflix, which employs NN AI to analyze user behavior and suggest personalizedcontent.

3.3 Applications of AI in the Digital Economic Era

This section presents the most notable applications of AI in the digital generation,including chatbots and virtual assistants, image and speech recognition, predictiveanalytics, autonomous vehicles, and healthcare, among others The following are thetop fifteen applications of AI that are already changing how we interact with technologyand the world around us.

3.3.1 Personalized Marketing

AI is critical in forming and applying personalized marketing by collecting and analyzingcustomer data, generating personalized content, and automating marketing delivery AIcan predict customer behavior through predictive analytics, optimize marketingmessages in real-time, and automate their delivery across multiple channels Using AIto collect and analyze customer data, businesses can create personalized marketingmessages that resonate with individual customers, increasing customersatisfaction, loyalty, and revenue (Hasin et al., 2022) The use of AI in personalizedmarketing has made it possible for businesses to deliver relevant messages at a higherscale, resulting in more effective marketing campaigns and better customerexperiences.

3.3.2 Image and Speech Recognition

AI plays a vital role in applying and forming image and speech recognition by usingdeep learning algorithms to analyze and interpret visual and auditory data Imagerecognition involves using AI to identify objects, faces, and scenes in images andvideos, while speech recognition uses AI to transcribe spoken words into text (Hazzanet al., 2022) AI algorithms can learn to recognize patterns in data and use thisknowledge to improve their accuracy over time Image and speech recognition havenumerous applications, including automated image tagging and classification, facialrecognition for security and authentication, and speech-to-text transcriptions foraccessibility and transcription purposes.

3.3.3 Virtual Assistants

AI plays a crucial role in forming and applying virtual assistants by enabling them tounderstand natural language, learn from user interactions, and respond to concurrentrequests (Yang et al., 2021) Virtual assistants are software applications that caninteract with users through text or voice commands, providing personalized assistanceand performing tasks on their behalf (Jun et al., 2021) AI algorithms can process andanalyze large amounts of data, including user behavior, preferences, and historicalinteractions, to better understand a user’s needs and intentions Virtual assistants canprovide intelligent recommendations, perform complex tasks, and automate routineprocesses, improving efficiency and productivity and developing more complex anduser-friendly technologies that transform how we interact with machines and enhancethe overall user experience.

3.3.4 Fraud Detection

AI plays a critical role in detecting and preventing fraud in digital generation, using AIalgorithms to detect and prevent fraudulent transactions, protect customer data, andprevent financial losses AI can simultaneously analyze massive volumes of data andidentify fraudulent activity more quickly and accurately than traditional methods (Shafiket al., 2020) In addition, ML algorithms can learn from historical data to detect new

fraud patterns and adjust their models accordingly The use of AI in fraud detection hasmade it possible to identify fraudulent activities more quickly and efficiently, improvingsecurity and trust in the digital economy (Zurdo et al., 2022).

AI has enabled the development of chatbots and computer programs that use naturallanguage processing to interact with users immediately It can be used for variouspurposes, including customer service, marketing, and sales (Shafik et al., 2020) Forinstance, e-commerce businesses use chatbots to provide personalized assistance tocustomers, helping them find products, answer their questions, and process orders(Yang et al., 2021) AI algorithms can learn from historical data to improve responsesand provide a more accurate and efficient customer experience Chatbots can also beused for lead generation and sales by engaging with customers and providing relevantoffers and recommendations (Zurdo et al., 2022) Using AI in chatbots has made itpossible to provide all-time assistance to customers, improving engagement and loyalty.

3.3.6 Predictive Maintenance

AI has enabled the development of predictive maintenance, a proactive maintenanceapproach that uses machine learning algorithms to predict when equipment is likely tofail and schedule maintenance before it occurs Like, industrial companies, use AI tomonitor the performance of their machinery and predict when maintenance is required,reducing downtime and improving productivity (Mondal & Tripathy, 2021) ML algorithmscan learn from anomalies that may indicate impending equipment failure and historicaldata to identify patterns, allowing businesses to take preventive measures before anyissues arise The use of AI in predictive maintenance has made it possible to preventunexpected equipment failures, improving safety and reducing the risk of productioninterruptions (Seo & Lee, 2020).

3.3.7 Autonomous Vehicles

Autonomous vehicles use a combination of sensors, ML algorithms, and advancedcomputer systems to navigate roads and make decisions in real-time For instance, self-driving cars use AI to detect other vehicles, pedestrians, and obstacles on the road andadjust their speed and direction accordingly (Shafik et al., 2022) ML algorithms canlearn from historical data to improve driving performance and make more accuratedecisions Autonomous vehicles have the potential to reduce accidents and improvetransportation efficiency by reducing traffic congestion and emissions By leveraging AI,autonomous vehicles can provide a safer and more efficient transportation system,improving the quality of life for people worldwide (Shafik et al., 2020) AI in autonomousvehicles has enabled mobility and convenience, revolutionizing the transportationindustry, as illustrated in Figure 3.1.