A REVIEW OF ARTIFICIAL INTELLIGENCE (AI) CHALLENGES AND FUTURE PROSPECTS OF EXPLAINABLE AI IN MAJOR FIELDS: A CASE STUDY OF NIGERIA.

Artificial intelligence (AI) has been used widely in essential fields such as energy, health, agriculture, finance etc. However, Artificial intelligence is still faced with social, ethical, legal, and technological challenges. It is important to know how these systems make their decisions while still achieving and implementing the benefits of AI. Explainable AI (XAI) is a technique that is used to explain how a machine made a decision. In this review, we discuss the challenges of AI and recommend XAI as a tool to solve the limitations of AI and suggest a human and conditions-based approach to challenges faced in the technology in Nigeria. This paper employs a narrative review to highlight problems that are limiting the use of AI in four important sectors of Nigeria: Health, Energy, Agriculture, and Finance, and suggest recommendations to solve the AI challenges. The review data was obtained from journals and researchers. We discuss Explainable AI (XAI) as a technique for solving challenges like trustworthiness, bias, lack of data, expertise, and confidence in using AI in major sectors. The paper focuses on the users, conditions, and challenges and recommends that humans and conditions be taken into consideration when building XAI systems.


INTRODUCTION
Research in Artificial Intelligence (AI) and Machine Learning (ML) has benefited human lives and has had positive impacts in various fields (Zhang & Lu, 2021).In addition to the vast existing contributions of AI, Artificial Intelligence still shows promising improvement which may impact human lives and contribute to the advancements of economies and societies in various sectors such as Health, transportation, security, and law.However, Artificial intelligence is still faced with social, ethical, legal, and technological challenges (Thiebes et al., 2021).Some issues of Artificial intelligence have brought about less trust in AI systems.Examples of these concerns with AI bias, AI systems may have problems of being biassed towards certain populations (Ghassemi et al., 2021;Obermeyer et al., 2019;Alikhademi et al., 2021), swapping faces of people in a video (deep fakes) (Adam et al.,2020), or clearview Ai, which could be used in monitoring people against their consent (Thiebes et al., 2021;Hill et al., 2020).
In Nigeria, Artificial intelligence has brought forth opportunities to contribute to areas such as Education (Abayomi, 2021;Sanusi et al., 2022), security (Falode et al., 2021), Energy (Mobayo et al., 2021), Health (Muhammad et al., 2021;Anazodo et al., 2022), however, several papers (Mobayo et al., 2021;Imhanyehor et al., 2021;Okwu et al., 2021) highlight a variety of novel challenges limiting the adoption of AI in Nigeria such as awareness, knowledge of field or system, adequate power supply, computing facilities and trusting of AI systems.These drawbacks could be solved with some technical conditions such as data, algorithms, and computing capabilities.These conditions have led to positive impacts in various areas of industries such as healthcare, retail, automotive, and finance in other parts of the world (Zhang & Lu, 2021).It is beneficial for these conditions to improve in Nigeria for better AI performance and to have some sort of explainability of AI decision systems to make these systems more trustworthy for use in Nigeria.
According to O'Neill (2021), in 2021, about 23.36 percent of Nigeria's GDP came from agriculture, industry such as manufacturing, processing, and transforming of goods contributed to 31.41 percent and 43.79% came from the services sector such as IT and banking.Oil is an important contribution to Nigeria.Between 2001 and 2010, Nigeria was among the countries with the highest GDP growth worldwide due to oil, though prices caused a growth slump (O'neill, 2021).
The services sector is growing due to the rise in the relocation of people from rural areas to urban areas (O'neill,2021).
Nigeria has the highest concentration of medical doctors but this is still not enough for the populous country of over 200 million people.The government contribution is lower than private contributions to health (Sasu, 2022).The health sector only has an investment of about 3 percent of Nigeria's GDP (Sasu, 2022).95% of Nigeria's foreign exchange earnings and 80% of its budgetary revenues are provided by the oil sector (The World Bank, n.d.).Artificial Intelligence can be described as the simulation of the human mind to make computers think and act like humans by performing tasks like learning and problem-solving (Zhang & Lu, 2021).Machine Learning builds systems that improve through data and experience and has been used in the advancement of various fields such as autonomous systems, Natural language processing, computer vision, and medical fields (Jordan & Mitchell, 2015).AI and ML are still being researched to achieve better performance.It is necessary to mitigate the bias and unfairness in data for ethical reasons and to gain users' trust in AI systems (Toreini et al., 2020) (Zhang & Lu, 2021;Ghassemi et al., 2021;Alikhademi et al., 2021).XAI is essentially Machine learning and AI technologies that provide humanunderstandable justification for their decision (Gunning et. al, 2019).
In the open literature, limitations of technology, and concepts of AI and their applications have been studied in some fields, but in the implementations and use of AI, research in Nigeria is limited.There is an issue connecting AI and its users and the conditions faced.The research is limited to implementing AI and is not focused on the user's view of AI.This study checks the drawbacks of technology and AI in Nigeria and suggests the use of XAI to curb issues.In Nigeria, basic techniques such as Management Systems are used in various sectors, However, there is a lack of use of AI technologies and data, which could help in the advancements of these sectors.These issues include a lack of skilled professionals, privacy, lack of trust, bias, lack of data, and other technologies.We need to use data obtained from Nigerian people or facilities to prevent bias and XAI to solve the issue of trust or lack of skill.This paper highlights the need for XAI and suggests how it can help four important sectors in Nigeria -Health, Finance, Energy, and Agriculture.AI and XAI are suggested to curb some of the issues faced in those sectors.Explaining how the system came to its decision could give better confidence in diagnosis, debugging, and verification, thereby creating more trustworthy systems.

AIMS AND OBJECTIVES
This study aims to highlight the prospects of AI in Nigeria.
The following are the objectives of the study:

MAJOR CONTRIBUTIONS
This study aims to highlight the prospects of AI in Nigeria.
The following are the major contributions of the study: 1. To review the challenges of Artificial Intelligence 2. To show the limitations of AI and technology in Nigeria 3. To understand why AI is not used in Nigeria 4. To suggest the use of XAI as a tool for the advancement of AI in Nigeria 5. To show how each of the limitations in Nigeria could be solved using XAI

PURPOSE OF THIS RESEARCH
The purpose of this research is to find ways explainable AI could aid major fields in Nigeria

NOVELTY
In this paper, we contribute to Academia in Nigeria by focusing on both the environment and the users in Nigeria.We review the technology to them being solved by XAI, We analyse the general challenges and suggest a method that could be implemented to suit Nigeria's condition.

BACKGROUND OF LITERATURE
This review analyses the drawbacks of Artificial Intelligence and reviews Explainable Artificial Intelligence as a way to curb these issues.Several researches have been carried out in Explainable Artificial Intelligence (XAI), however, there is no standard method for evaluating or comparing XAI methods.Hence this research reviews the issues of AI and XAI and highlights the need for XAI in various fields field

EXPLAINABLE ARTIFICIAL INTELLIGENCE
There are various reasons for XAI which include trustworthiness, causality, transferability, privacy, awareness and fairness.

METHODS
Explainable Artificial Intelligence is typically divided into two types.The first type Inherent explainability, is where models can be inherently interpretable from their design, for example, linear regression models where as one variable goes up another variable moves up.The second type, post-hoc explainability, is where models can be interpreted from

HEALTHCARE
The lack of Skilled professionals and required systems for progress and growth is poor across the board in the health sector in Nigeria (Okoroafor, 2021).Okoroafor (2019) noted that rural areas are understaffed in the health sector due to poor social amenities like water, electricity, telecommunications, security, infrastructure, and working conditions.
In Nigeria, various challenges affect the healthcare sector.Okolie et al (2022) et al., 2015).Some challenges facing access to Sexual and Reproductive Health services are distance to health facilities and opening hours, financial problems, and community and religious beliefs (Nmadu et al., 2020).
AI has been used to improve classification and prediction in healthcare (Verma et.al., 2022).Deep learning, where neural networks can classify medical images from raw data has achieved significant progress.XAI techniques such as Bayesian filtering and salience maps have been used to achieve transparency in AI models (Folke et al., 2021).(Saraswat et.al., 2022) outlined the basics of XAI, associated metrics, and use-case applications for classifying and segmenting covid-19 patients.Their study presented the integration of FL and XAI for decentralised healthcare setups and showed the benefits of AI in health setups.Heat maps or saliency map is a commonly used form of post-hoc explainability (Rajpurkar et. al, 2021) that highlighted the affected area with a saliency map, however, the hottest part, which should be the most useful part, showed both useful and non-useful information and does not explicitly reveal the useful part of that area.It is difficult to measure the ability of the explanatory technique because heatmaps may appear believable even with untrained models.Loh et al., (2022) reviewed research on various XAI techniques and discovered that abnormality detection in 1d biosignals and identification of clinical notes key text required more attention in XAI.(Giuste et. al., 2022) carried out a systematic review of XAI for combating pandemics.In their work, they reviewed approaches in XAI that increase the adoption of AI based on lessons learned from COVID-19.They highlighted that due to the risk-averse nature of clinical information, developers need to know how an AI decision system came to its decision.(Giuste et. al., 2022) found that confidence due to unbiased data and bias detection and pattern-discovery in the data were of benefit and will improve patient care.The papers may have achieved XAI, which is hard to measure as there aren't quantifiable means.However, there is a gap in relation to users.Users Should be involved in the development process of youtube the problem to specific users.(Marvin, 2022) developed lattice-based Machine Learning models for explainable prediction of fertility treatment intervention outcomes their research still will be greatly improved if there was research about health centre practices related to neonatal intensive care unit admission to recognise hidden parts to reduce emergency cases.(Khan et. al., 2022) noted the scarcity of professionals, awareness, and tools in point-of-care settings is one of the main causes of maternal and newborn deaths and noted a need for explainability in AI to solve ethical issues.

ENERGY
The Nigerian oil and gas industry has been affected by huge costs while undergoing the exploration and production of fossil fuels (Nkwoji, 2021).Due to oil spills, The Nigerian economy experienced a loss of approximately 3,928,260,196 naira in revenue between 1984-2012.This estimation excludes environmental impact, third-party, and remediation costs (Osuagwu & Olaifa, 2018;Chinonyerem et al., 2017).An average of 700 spills are recorded annually causing environmental degradation, while also having a negative economic impact (Ejiba et al., 2016).Mobayo et al. (2021) carried out a study and found that the major challenges of the energy sector in Nigeria are scarcity of skilled professionals, cellular technologies, outdated power system infrastructure, and the growing threat from cyber-attacks cellular technologies.Machine learning algorithms could be used to recognize patterns from data and could be used to identify anomalies.Top energy companies are moving to be data-directed companies.It is essential to help companies understand various performances, including the balance of demand and supply (Nisi et al., 2015).This requires a deep analysis of factors such as products, equipment, weather, days, holidays, and customers.It is also necessary to process data and identify the correlation between data features (Zainab et al., 2021;Nisi et al., 2015).XAI has been used in energy and power systems.(Santos et. al., 2022) used an LSTM model in grid disturbances and carried out the combination of DeepLift andSHap for explainability which provided insights into the misclassifications.(Machlev et.al, 2021) showed the possibility of evaluating explainability in XAI technique and deep learning classifiers and showed how some of the strongest features are not directly related to disturbance when explained with GRAD-CAM.Most XAI in energy use SHAP and LIME, and use traditional Machine learning algorithms and not deep learning (Machlev et. al.,2022).(Machlev et. al.,2022) Highlighted challenges that need to be addressed when implementing XAI techniques in the energy and power systems domain.These include trade-offs, users, standardisation, recommendations, security, and evaluation metrics.

AGRICULTURE
Agriculture is one of the most vital means of reducing poverty in Nigeria (Omodero, 2021).Omodero (2021) researched that participation of youth in agriculture as a primary profession could increase the per capita farm income by ₦31,301.22 -₦32,150.526and could reduce poverty by 17% (Osabohein, 2021).Osabohein, (2021) discovered that some farmers do not have enough skills to control the pests ruining farm crops and suggested that the government should supply the required facilities such as training, huge markets, huge markets, modern farming gadget, adequate power supply, and storage spaces to improve agriculture in Nigeria (Omodero, 2021).The cost of farming in Nigeria has increased and this has led to reduced profitability (Osabohein, 2021).They suggested that farmers open tier markets emotionally.The increased cost has an impact on farmers' decision to produce for international markets (Osabohein, 2021).
Global food production should be increased by an estimated 60-70% to feed the increasing population of 9 billion people by 2050 (Rockstrom et al., 2017).Using ML in four sectors (reproduction, production, processing, and distribution) in Agriculture could be very beneficial (Ahumada & Villabos, 2009).ML could be used to predict crop yield, soil properties, disease detection, weather prediction, production planning, distribution cluster, storage, transportation, and consumer analysis (Ben & Hanana, 2021).Ryo (2022)  for data analysis in agriculture.The methods could be applied to any machine learning algorithms after training hence they were post-hoc and model-agnostic.The LIME method was used to show that local variables differed from global ones because of certain conditions.XAI involves the understanding of how organisations deploy AI systems and the user interface in human-Ai joint decision-making (Cartolano et al.,2022).

BANKING AND FINANCE
It was found that most banks in Nigeria utilise chatbots that may enhance engagement with customers The most frequently used platform for this was WhatsApp was the most frequently used platform.The chatbots were not able to perform as well when out of their predefined path.The chatbots used English and not any of the local languages.
Managers know the promise of AI and are open to establishing AI in business banking but note the issues hindering the adoption of AI (Mogaji & Nguyen, 2022).However, a study by Mogaji and Nguyen (2022), emphasises that banks must understand the objectives of the business, the resources available, and customer needs.Banks should train managers and ensure regulators are involved in the development and inform consumers about AI possibilities.
The utilisation of AI and ML in financial services is transforming industries and societies.Many Financial firms, from financial technology (FinTech) service providers, to fund management firms, to retail and investment banks, are employing Machine Learning expertise (Goodell et al., 2021).
There has not been much study on XAI in Finance (Černevičienė, 2022).The first XAI study in finance was conducted in 2020 by Bussmann et.al (2020), Ariza-Garzón (2020) and Garmegna et.et. al., 2022).There is more concentration on informativeness and confidence and trustability and less focus on accessibility, causality, and privacy awareness (Arrieta et al., 2019;Weber et. al., 2022 ).Weber et al. (2022) summarised XAI models, which may be under-researched more than others for further research.

ISSUES OF ARTIFICIAL INTELLIGENCE
Various issues of XAI have been examined in explainable Artificial intelligence.One challenge is that there is no standard for evaluating Artificial Intelligence.One major drawback of XAI is the lack of user involvement in XAI processes.Humans should be involved in the process of XAI for particular scenarios.
Some issues limiting the growth of AI, which affect the Health, Energy, Agriculture, and Finance fields are: • Data reliability, data uniformity, and Data Volume -: Due to equipment or network failure, data may not be complete and lead to erroneous results.The data may also be noisy or only focused on one aspect.Data could also be collected from multiple sources or may be limited.Combining and finding data may be also difficult (Chen et al., 2015;Sen et al., 2021).Obtaining health data is difficult due to various factors, Electronic Health Records could be stored in various formats making them incompatible with other platforms due to interoperability issues.EHRs make use of different terminologies, symbols, and coding values (Ehrenstein, 2019).

Mohammed & Shehu, 2023 OJPS 4(1) | 08
• Privacy is also an issue with obtaining health records as patients may not consent for their data to be shared (Ehrenstein, 2019).
• Standardisation-Technical and code formats may be stored in various formats such as XML, JSON, etc (Brewseter et al., 2017).Research needs to be carried out to promote standardisation to increase interoperability among smart systems (Friha et al., 2021) • Bias and unfairness in the algorithm/data -if the training data for a model is biased, it may lead to unreliable results.Some records may only contain information from a certain race, gender, or socio-economic factors.
Data from urban areas will provide less accurate results when used on people from rural areas.Algorithms should be trained using more diverse data.An issue limiting AI research in healthcare is the lack of standardised methodologies and the use of only historical data.Patients' data should be studied over time.
• Regulation issues: It is important to set out the legal and regulatory rules for governing the management and control of the data.(Parasuraman, 2021).Countries can have different regulations with regard to service facilities such as technological challenges, data protection, and security.
• Market issues: There is a lack of sufficient awareness of smart devices, especially in rural areas.
• Security: Smart devices may be tampered with (Haseeb et al., 2020).Demestichas (2020) highlighted various security breach cases namely SQL injection attacks, Data Theft, and so on at various IoT layers.Insecurity of Data and insufficient information are part of the main and complex issues of systems (Guo et al., 2019;Zhao et al., 2019;Zhao et al., 2013).It is vital to have means of preventing cyber-attacks because of the cost.
Installing digital infrastructure in energy costs approximately $30 billion a year (Fickling, 2019).
• Lack of skills and practical expertise: This adds to the slow growth of AI among decision-makers.Cost of faults is high, which makes it difficult for companies to adopt the use of these new technologies (Ahmad et al., 2021).Using AI in some energy could be complex and challenging for operating grids (Puri et al., 2019).
• Outdated system infrastructure: There may be available data in the AI sector, however, data is disorganised and systems may be outdated (Ahmad et al., 2021).
• Black boxes: Consumers may not understand the internal functions of AI-based applications (Ahmad et al., 2021).
• Complexity: Some systems may e complex and have high dimensionality of data which could be a challenge in systems (Tang et al., 2018)

LIMITATIONS OF THE CURRENT STUDY
The current work shows a limitation that does not focus on the user or is not country-specific.Most studies were tested out by developers and not users, some were carried out on non-Nigerian data, which may cause bias.Some studies are also suited for research and summarise XAI methods, some focus on confidence.However, none focus on challenges and conditions suited to Nigeria.

OBSERVATIONS AND RECOMMENDATIONS
Mohammed & Shehu, 2023 OJPS 4(1) |09 We observed various challenges in AI in Nigeria.For example, In Abuja, Nigeria, we carried out a task of obtaining clinical data for cervical and Ovarian cancer for Machine Learning research.However, we were not able to obtain the data for research due to the unavailability of data.The records were either deleted, not stored for over a year, not annotated, or not approved for distribution.Data for cervical cancer was available on the internet but could not be used to make a reliable output for women in Nigeria due to Nigeria being underdeveloped while the data retrieved was uploaded from a developed country.
We researched statistical data contributing to the growth of Nigeria and hindering the well-being of Nigerians.The following fields were analysed: • Health: We discovered that this vital field lacked factors like practical expertise.
• Energy: This is an important field in Nigeria.Oil contributes significantly to the GDP.
• Agriculture: There is a need to improve agricultural practices as the human population keeps growing, • Banking and Finance: The Services sector contributes to almost half of Nigeria's GDP.Banking and IT are examples of these services.

RECOMMENDATIONS FOR NIGERIA
We suggest that when building XAI systems for Nigeria, both humans and conditions be taken into account.Using humans would provide additional feedback and help in debugging the system.Using conditions will also give feedback based on those conditions which may be unique to Nigeria.XAI should provide explanations and justify systems based on conditions.

RECOMMENDATIONS FOR THE HEALTH SECTOR IN NIGERIA
Based on our research, we suggest that the standard medical terminologies should be used with data retrieved from Nigeria.We suggest data should be stored based on ethical principles putting into mind privacy and bias.Hospitals should try to work with researchers to try to solve healthcare challenges with AI technologies.To mitigate bias, the data stored should include records of various populations (such as socio-economic factors) to provide more reliable results -this can also show how people from a certain region or socio-economic factor are affected more by particular disease and AI technologies could be used to highlight these populations for further actions by aiding bodies.We also suggest that Explainability should work with these expert systems to give end-users of the system (i,e radiologist or patients) more trust and confidence in the decisions.We suggest a baseline for explainability be defined.

RECOMMENDATIONS FOR THE ENERGY SECTOR IN NIGERIA
Based on our research, we suggest the following for Energy in Nigeria: There is data for the energy sector, however, due to the lack of expertise, we suggest XAI should be explored.Having some sort of explainability could help the workers who lack experience by suggesting the system has output this due to this, so they can make more confident decisions.

RECOMMENDATIONS FOR THE AGRICULTURAL SECTOR IN NIGERIA
Studies show youth are positive to participate in agriculture and show that this can reduce poverty and increase per capita income.However, the youth may not know about farming there for XAI should work alongside ML applications to make the new farmers more confident in what the systems suggest when asked about a problem.
Different foods are grown across various parts of the world.So, could the agricultural data used in another part of the world be used elsewhere?Would the decisions or recommendations from the AI system be reliable?This might give an issue of bias.Having some Explainability may back the decisions better.

RECOMMENDATIONS FOR THE FINANCIAL SECTOR IN NIGERIA
We discovered AI is mainly used for chatbots in Nigeria, and suggest that AI applications be implemented in Nigeria to help with risk assessment and monitoring, cyber security, and advisory services.Workers should be trained and made aware of AI uses in Nigeria.We also suggest the use of Explainable Artificial Intelligence as an aid to prevent bias, and aid in accountability and transparency.

CONCLUSION
Artificial intelligence has transformed various industries and societies.AI has had a positive impact on human lives and well-being.However, AI also comes with challenges that need to be tackled.Due to the impact of AI decisions on human lives, users of an AI system should be given explanations of certain decisions to be more confident in decisions.The current research does not focus on conditions in Nigeria.The focus is more on confidence, and methods and not on conditions or challenges.There is no standard way to evaluate AI, hence involving humans and using conditions pertaining to Nigeria works best.

Mohammed & Shehu, 2023 OJPS 4(1) | 11
The global issues of AI include the following: • Localisation: can the AI system work diversely across various regions, demographics, cultures, and languages?It is important to keep the numerous markets in mind while building models so they can adapt to users' experiences.
• Difficulty in data acquisition: We explained how difficult it was to acquire data in Nigeria due to unavailability, unlabeled/not annotated, not being recorded, or unapproved.We also explained that there may be a problem of bias if we were to use data available on the internet because of various factors like race, gender, and socio-economic factors.
• Accountability, if a problem goes wrong, one would not know who to blame, XAI can give users an explanation for example "If you make X it will give you Y due to Z" that as we can see why the system is telling the user to pick 'X', and the user will only pick X if they are satisfied with the explanation.
• Lack of expertise: if one is not an expert in a certain field and the expert system gives a recommendation, the system can say "this is X due to the Y being at Z and as big W do X so it does not become faulty due to Y" then the practitioner may know how and why the algorithm has arrived at a specific decision in following user's made a decision • Trustworthiness: In health for example, if neither the doctors nor patients do not trust the system, they will unlikely not follow the prognosis or diagnosis output by the system.
With these issues, it is believed that XAI should be explored in Nigeria to curb the issues discussed above.XAI should aim to give accurate, fair, transparent output that will make us understand why a system came to a particular decision.So, if there is a problem with data, we want the system to say it is bringing this result due to a certain factor.This way we can see why and spot if the explanation is satisfactory to the end-user of the system.
AI is still being explored in Nigeria and still has many areas that need exploration and implementation.Some issues limiting AI in Nigeria are.Only when clients are informed, convinced, and confident in AI models can they be effective in making accurate predictions.Consumers would prefer to use a system only when data is secured The issues with AI in Nigeria are lack of data -We found out that it was difficult to acquire data as it was difficult to obtain from data centers, not annotated or not available, Standardisation, Lack of Skilled professionals, Infrastructure, Computational capabilities, Awareness, lack of confidence, Security, Awareness.
We recommend an approach that takes into account users and conditions.Our suggestions for AI in Nigeria are: • A human-in-the-loop approach should be used, to obtain satisfaction in XAI • Conditions should be taken into consideration to build a system that fits the problem.
• Organisations should cooperate in research, this includes storing data ethically, annotating data (giving it correct labels), and sharing data ethically for research purposes.
• Organisations should have a standard set of terms to use in research.
• It is also important to get data across various populations to curb the problem of bias.and to make data more inclusive.For example, with the chatbots, we see that most applications are English based but expansion to other languages such as Hausa, Yoruba, and Igbo should be carried out.
Mohammed & Shehu, 2023 OJPS 4(1) | 12 • More research and implementation should be carried out in Artificial Intelligence in Nigeria to solve the issues highlighted in this paper.
• More research should be carried out in XAI -though research in XAI still has more room for exploration, XAI can still be used in Nigeria using the influence and visualisation methods that have been implemented elsewhere.
• Nigeria can also join in the research for DL models in Explainable Artificial Intelligence.

CONFLICT OF INTEREST
We declare that there is no known conflict of interest that could influence our work.
complex models and high-dimensional data such as analysing images, sound, and text.Post-hoc explainability includes saliency maps, locally interpretable model-agnostic explanations (LIME), and Shapley additive explanations (SHAP).

Figure 3 :
Figure 3: Chronology of the development of Explanatory methods.Source: Holzinger et al. (2022) used the global data set of maize crop yield al (2021) Due to the difficulty in evaluating XAI models, Černevičienė (2022) suggested a Multi-Criteria Decision system (MCDM) to combine different mechanisms based on system references and adjunction policy in finance, however, they still need more work in integrating AI and MCDMS.Research in XAI in finance is very dispersed in application areas and methods (Weber

Figure 4 :
Figure 4: Causability and Explainability in Human-AI Interaction Source: Shin (2021) also observed various factors, stating the importance of infrastructure, adequate knowledge, and attitudes, cost of screening, anxiety about positive results, and low-risk awareness among Female Health workers.According to the World Health Organization (WHO) (2019), Nigeria accounts for 20% of global maternal deaths.Sexual and reproductive health is low, especially in rural areas and among the youth (Denmo