Project – Explainable Artificial Intelligence and User Trust in Automated Decision-Support Systems: A Study of Selected Nigerian FinTech Applications

CHAPTER ONE

INTRODUCTION

1.1 Background to the Study

Artificial Intelligence (AI) has become an increasingly important component of modern financial technology, changing how financial institutions and FinTech companies process information, assess risks, detect fraud, personalize services, and make operational decisions. Unlike conventional software that follows explicitly programmed rules, contemporary AI systems can learn patterns from large datasets and use those patterns to generate predictions or recommendations. In financial services, these capabilities are particularly relevant because decisions such as credit assessment, fraud detection, customer profiling, transaction monitoring, and risk classification often require the processing of large volumes of data within very short periods. The increasing use of complex machine-learning models, however, has created concerns about how users can understand and evaluate decisions generated by systems whose internal reasoning may be difficult to interpret. Arrieta et al. (2020) explain that the increasing complexity of machine-learning models has made explainability an important requirement for responsible AI, particularly where decisions have significant consequences for individuals. Similarly, Weber et al. (2024) identify explainability as an important issue in financial applications of AI because complex models may produce useful predictions while remaining difficult for users and decision-makers to understand.

Explainable Artificial Intelligence (XAI) has emerged partly in response to the opacity associated with many modern AI and machine-learning systems. XAI refers broadly to approaches that make the reasoning, factors, or processes underlying an AI-generated outcome more understandable to human users. Explainability does not necessarily mean exposing every technical operation within an algorithm; rather, it involves providing meaningful information that enables relevant users to understand why a particular recommendation, prediction, classification, or decision was produced. Miller (2019) argues that explanations in AI should be considered from the perspective of how people understand explanations in social and cognitive contexts, rather than treating explanation solely as a technical problem. Rai (2020) similarly observes that explainable AI can help move AI systems from opaque “black boxes” toward systems whose decision processes are more understandable and accountable. Consequently, explainability has become increasingly relevant in environments where people must decide whether to accept, question, or act upon AI-generated outcomes.

The importance of explainability becomes even more pronounced when AI is used for automated or semi-automated financial decision-support. FinTech applications increasingly employ algorithmic systems for activities such as credit scoring, fraud detection, customer verification, transaction monitoring, financial recommendations, and other risk-related decisions. Such applications can improve processing speed and potentially increase the consistency and scalability of financial services. Nevertheless, the use of AI in financial decision-making also raises concerns about fairness, bias, privacy, accountability, and the ability of affected users to understand why particular outcomes were produced. Tigges et al. (2024), in their study of AI and alternative data in FinTech lending, note that AI-based credit assessment can improve risk management and expand access to credit, while simultaneously creating concerns relating to consent, data quality, transparency, traceability, bias, discrimination, and data misuse. Weber et al. (2024) likewise show that explainability has become a significant research concern across financial applications of AI.

User trust is therefore a central consideration in the successful deployment of AI-enabled financial services. Trust in this context refers to the willingness of users to rely on an AI-enabled system and accept its outputs as sufficiently dependable for action or continued use. In FinTech, trust may influence whether customers are willing to provide personal information, follow automated recommendations, accept digital financial decisions, continue using a particular application, or adopt AI-enabled financial services. Zarifis and Cheng (2022) developed a model of trust in FinTech and found that trust is influenced by individual, social, organizational, and technological factors, including trust in AI and related technologies. Similarly, Akinwale and Kyari (2022), studying FinTech users in Lagos State, found that service trust had a significant positive influence on users’ attitudes toward FinTech services, while users’ attitudes significantly influenced adoption. These findings indicate that technological functionality alone may not be sufficient to encourage sustained use of AI-enabled financial platforms; users must also have confidence in the systems and organizations behind them.

Explainability may provide an important mechanism through which confidence in automated decisions can be developed. When users receive understandable information about why an AI system produced a particular outcome, they may be better positioned to evaluate the reliability, relevance, and fairness of that outcome. Ribeiro, Singh, and Guestrin (2016) demonstrated the importance of explanations for helping people assess whether individual machine-learning predictions should be trusted. In a related empirical study, Shin (2021) found that explainability and the ability of users to understand explanations can influence perceptions and trust in AI-generated outcomes. However, explainability does not automatically guarantee trust because poorly designed, overly technical, incomplete, or confusing explanations may fail to assist users. Miller (2019) therefore emphasizes the importance of designing explanations around human understanding, while Arrieta et al. (2020) identify the audience for whom an explanation is intended as an important consideration in explainable AI.

The Nigerian FinTech environment provides an important setting for examining the relationship between explainability and trust because financial technology has become increasingly integrated into everyday financial activities. Nigerian users interact with digital financial applications for payments, transfers, lending, savings, investment, account management, and other financial services, while technology providers increasingly employ automated tools to improve operational efficiency and manage financial risks. The Central Bank of Nigeria (CBN) recognizes the importance of transparency, accountability, sound internal controls, consumer protection, and effective supervision in the FinTech ecosystem. Recent CBN reporting also indicates that AI adoption within Nigeria’s financial sector is growing, although regulatory clarity, technical capacity, infrastructure, and other constraints remain relevant considerations. In addition, the Nigeria Data Protection Act 2023 recognizes automated decision-making and establishes a legal framework for the protection of personal data, making issues surrounding responsible processing and automated decisions increasingly important to Nigerian digital financial services.

Despite the growing relevance of AI, explainability, and trust in financial technology, there remains a need for empirical research that examines these issues specifically from the perspective of users of Nigerian FinTech applications. Existing studies have examined FinTech adoption, digital financial transactions, consumer trust, AI applications in finance, and the technical development of explainable AI, but these areas are often studied separately. For example, Kajol, Singh, and Paul (2022) identified trust as one of the most frequently cited motivators of digital financial transaction adoption, while perceived risk and privacy concerns remain important barriers. Nigerian research has also identified trust, security, and technological factors as important to FinTech adoption, but comparatively less attention has been given to whether the explainability of AI-generated decisions directly contributes to users’ trust in automated decision-support systems within Nigerian FinTech applications. This creates a significant research gap, particularly as AI becomes more embedded in financial decisions that may affect users’ access to credit, transaction security, financial recommendations, and other services. Therefore, this study seeks to examine Explainable Artificial Intelligence and User Trust in Automated Decision-Support Systems: A Study of Selected Nigerian FinTech Applications.

1.2 Statement of the Problem

The increasing deployment of AI in financial technology has created opportunities for faster, data-driven, and potentially more efficient financial decision-making. However, the complexity of AI systems can make it difficult for ordinary users to understand how particular outcomes are reached. In FinTech applications, a user may receive a credit decision, fraud alert, transaction restriction, financial recommendation, or risk assessment without knowing the major factors that contributed to that outcome. This creates an important problem because users may find it difficult to determine whether an automated decision is accurate, fair, appropriate, or biased. Arrieta et al. (2020) identify opacity and limited interpretability as major challenges to the responsible deployment of complex AI models, while Weber et al. (2024) emphasize that explainability is particularly important in finance because AI-generated decisions can have significant consequences for individuals and organizations.

A second problem concerns the relationship between explainability and user trust. Financial decisions are often sensitive because they may directly affect individuals’ money, access to credit, transaction security, or financial opportunities. When an automated system makes a decision that users do not understand, users may perceive the system as unreliable, unfair, or unsafe, even where the underlying model performs effectively. Conversely, explanations that are understandable and relevant may help users evaluate AI outputs and develop greater confidence in the system. Ribeiro et al. (2016) demonstrate the usefulness of explanations in helping users determine whether machine-learning predictions should be trusted, while Shin (2021) provides evidence that explainability and users’ ability to understand explanations are connected with trust and attitudes toward AI. The problem, therefore, is not simply whether Nigerian FinTech companies use AI, but whether users understand and trust the decisions produced by such systems.

A third problem is that Nigerian FinTech users operate within an environment where trust, security, privacy, and perceived risk are already important determinants of digital financial behaviour. Akinwale and Kyari (2022), for example, found that service trust significantly influenced attitudes toward FinTech services among users in Lagos State. Mogaji and Nguyen (2022) similarly identified fraud, security, information privacy, and other concerns as important issues surrounding FinTech and mobile money in Nigeria. These concerns become more complicated when financial decisions are increasingly supported or generated by AI because users may not know what data are being considered, how those data are interpreted, or why the system has produced a particular result. The Nigerian Data Protection Act 2023 also makes automated decision-making and personal-data protection important regulatory considerations, reinforcing the need to examine how users perceive automated decisions and the transparency surrounding them.

A fourth problem is the relative scarcity of empirical evidence specifically examining how explainability affects user trust in AI-based automated decision-support systems within Nigerian FinTech applications. Existing international research has established important relationships among explainability, transparency, trust, fairness, and AI acceptance, while Nigerian studies have largely focused on FinTech adoption, digital payments, cybersecurity, AI applications, or general customer trust. Recent Nigerian evidence indicates increasing interest in AI applications in fraud detection, credit assessment, and financial services, but there remains a need to examine the user’s perspective of the explanations accompanying AI-supported decisions. The gap is important because a technically accurate AI system may still encounter resistance if users do not understand or trust its outputs. Consequently, this study is designed to investigate the extent to which explainability of AI-based automated decision-support systems influences user trust in selected Nigerian FinTech applications.

1.3 Purpose of the Study

The main purpose of this study is to examine the relationship between Explainable Artificial Intelligence and user trust in automated decision-support systems among users of selected Nigerian FinTech applications.

The specific objectives are to:

1. examine the extent to which Explainable Artificial Intelligence is incorporated into automated decision-support systems used by selected Nigerian FinTech applications;
2. determine the level of user trust in AI-supported decision-making among users of selected Nigerian FinTech applications;
3. examine users’ perceptions of the clarity and understandability of explanations provided for AI-generated decisions;
4. determine the influence of perceived transparency of AI decisions on user trust

1.4 Research Questions

The following research questions will guide the study:

1. To what extent is Explainable Artificial Intelligence incorporated into automated decision-support systems used by selected Nigerian FinTech applications?
2. What is the level of user trust in AI-supported decision-making among users of selected Nigerian FinTech applications?
3. How do users perceive the clarity and understandability of explanations provided for AI-generated decisions?
4. To what extent does perceived transparency of AI decisions influence user trust?

1.5 Research Hypothesis

The following null hypothesis will be tested at the 0.05 level of significance:

H₀: There is no significant relationship between Explainable Artificial Intelligence and user trust in automated decision-support systems among users of selected Nigerian FinTech applications.

1.6 Significance of the Study

The study will be significant to FinTech companies because it will provide empirical information on how users perceive AI-supported decisions and the extent to which explainability may contribute to confidence in their platforms. The findings may assist FinTech developers and managers in designing decision-support systems that provide clearer, more relevant, and user-oriented explanations for automated outcomes.

The study will also be significant to users of FinTech applications. By examining transparency, explainability, fairness, and trust, the study may contribute to greater awareness of how automated decisions affect users. It may also encourage the development of financial technologies that communicate AI-generated outcomes in ways that users can understand and evaluate.

The findings will be useful to regulators and policymakers, particularly institutions concerned with financial technology, consumer protection, data protection, and digital financial services. The study may provide evidence that can support policies concerning transparency, accountability, responsible AI, automated decision-making, and consumer protection within Nigeria’s evolving FinTech ecosystem. This is particularly relevant because the CBN emphasizes consumer confidence, transparency, accountability, and sound practices in the financial and FinTech system, while the Nigeria Data Protection Act 2023 provides a framework for responsible processing of personal data and recognizes automated decision-making.

The study will be valuable to AI developers, data scientists, and information-systems professionals because it will provide a user-oriented perspective on explainability. Rather than focusing only on the technical accuracy of machine-learning models, the study will draw attention to whether explanations are understandable and capable of supporting appropriate levels of user confidence.

Finally, the study will contribute to academic research in information systems, artificial intelligence, FinTech, technology adoption, and digital trust. It will provide a Nigerian context for examining the interaction between explainability and trust and may serve as a basis for further studies on responsible AI and automated decision-making in developing economies.

1.7 Scope of the Study

The study focuses on Explainable Artificial Intelligence and user trust in automated decision-support systems among users of selected Nigerian FinTech applications.

Conceptually, the study will focus on Explainable Artificial Intelligence as the major independent variable and user trust as the dependent variable. The dimensions of explainability considered in the study will include perceived transparency, understandability, clarity of explanations, perceived fairness, and accountability of AI-supported decisions. User trust will focus on users’ confidence in the reliability, dependability, fairness, security, and appropriateness of AI-supported decisions.

The study will concentrate on FinTech applications that use or are perceived by users to use automated or AI-supported decision processes, including areas such as digital lending and credit assessment, fraud detection, transaction monitoring, customer-service automation, financial recommendations, and related financial decision-support activities.

Geographically, the study is limited to selected Nigerian FinTech applications and their users in Nigeria. The particular FinTech applications and respondents will be determined according to the sampling criteria established in the methodology chapter.

1.8 Operational Definition of Terms

Artificial Intelligence (AI): The capacity of computer-based systems to perform tasks that normally require aspects of human intelligence, including learning, prediction, pattern recognition, reasoning, and decision support.

Automated Decision-Support System: A computerized system that uses algorithms, data, artificial intelligence, or machine learning to generate recommendations, predictions, classifications, alerts, or other information intended to support or influence decisions.

Explainable Artificial Intelligence (XAI): Methods, techniques, and system features that make the reasoning or factors underlying AI-generated outputs understandable to relevant human users.

User Trust: The willingness or confidence of a user to rely on an AI-enabled financial system and accept its outputs as sufficiently dependable, accurate, fair, and appropriate for use.

FinTech: The application of technology to the provision, delivery, management, or improvement of financial products and services.

FinTech Application: A digital platform, mobile application, web-based service, or technology-enabled financial product through which users access financial services such as payments, lending, transfers, savings, investments, or related services.

Transparency: The extent to which relevant information about an AI system’s operation, inputs, processes, or decisions is made understandable and accessible to users or other appropriate stakeholders.

Interpretability: The extent to which a human user can understand how an AI model or system arrives at a particular output.

Algorithmic Decision-Making: The use of computational rules, algorithms, statistical models, or AI systems to generate or influence decisions.

Perceived Fairness: The extent to which users believe that AI-supported decisions are impartial, non-discriminatory, and based on appropriate factors.

Accountability: The extent to which responsibility for the development, operation, monitoring, and consequences of AI-supported decisions can be identified and assigned to relevant persons or organizations.

Project – Explainable Artificial Intelligence and User Trust in Automated Decision-Support Systems: A Study of Selected Nigerian FinTech Applications