INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XIII September 2025
Special Issue on Emerging Paradigms in Computer Science and Technology
|
Page 1
www.rsisinternational.org
The Ethics of AI in Financial Planning: Bias, Transparency, and the
Role of Human Judgment
Dolapo Achimugu
1
, Chinaza Ukatu
1
, Arinze E. Anaege
2
1
College of William & Mary, Raymond A Mason School of Business, Williamsburg, Virginia US
2
Department of Accounting, Kingsley Ozumba Mbadiwe University, Ideato, Nigeria
DOI:
https://doi.org/10.51244/IJRSI.2025.1213CS001
Received: 10 September 2025; Accepted: 16 September 2025; Published: 15 October 2025
ABSTRACT
The fast-growing use of artificial intelligence (AI) has introduced new ethical issues in the financial services
sector. Robo-advisors, loan algorithms, and automated financial instruments now make choices that impact
people's lives substantially. These automated instruments may lack fairness, clarity, and human supervision.
Without adequate checks, they could generate discriminatory decisions or erode trust in financial institutions.
This paper sets forth a normative-ethical framework to help oversee the responsible use of AI in financial
planning. The study identified a novel framework known as the EFT Model, which has four pillars: Ethical
Design, Fairness, Transparency, and Human Oversight. The paper examines each principle in detail, illustrating
it with practical examples like discriminatory loan approvals and unclear investment recommendations. Roles
and accountability of key players such as developers, regulators, financial institutions, and customers are also
clearly identified. The paper harmonizes the framework with existing regulations like the EU AI Act, the GDPR
and discusses how it could help direct ethical design in practice. It also underlines the importance of conducting
additional research with the intention of testing and refining the model under real-world conditions.
Keywords: AI Ethics, Fintech, Ethical Framework, Human Oversight, Transparency, Fairness
INTRODUCTION
Financial services are being rapidly transformed with the use of Artificial Intelligence (AI). Among the most
significant applications to financial planning are AI-enabled products such as robo-advisors, scoring engines,
and algorithmic investment products (Sahu, 2024). These hold the promise of higher efficiency, scale, and
differentiated service offerings, often beyond human capacity in aggregating vast sets of financial information.
As investors and institutions rely increasingly on AI support, not only in investment decisions, but in insurance
planning, retirement, and in wealth management, its effects on financial outcomes keep growing (Vuković et al.,
2025). The rapidity with which AI has been introduced into financial planning, however, poses ethical issues.
Even as efficiency enhancement and cost reduction are normally the key driving forces underlying AI
introduction, bias, lack of explainability, and lack of human supervision are emerging risks (Černevičienė &
Kabašinskas, 2024). AI algorithms utilized in scoring, such as those employed in underwriting, are found with
tendencies to perpetuate existing social biases (Agarwal, 2024). In some cases, customers' loan applications are
rejected or offered at unfavorable terms. The lack of explainability limits the autonomy of the customer, reduces
trust, and creates gaps in accountability (Nallakaruppan et al., 2024).
Gladstone and Hundtofte (2023) noted that financial planning is an area where decisions make long-duration
impacts on people's financial well-being and safety. It involves a great deal of trust, ethical judgment, and
contextual awareness. These traits are not always transferable to AI systems, particularly black-box AI like deep
learning. The lesser role played by human advisors at key points in decisions has additionally raised an ethical
issue about offloading moral responsibility onto machines. While AI can facilitate decisions, the elimination of
human intervention in decisions involving high financial stakes becomes ethically problematic (Giarmoleo et al.,
2024). Despite rising awareness about these challenges, most of what has been written about AI in financial
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XIII September 2025
Special Issue on Emerging Paradigms in Computer Science and Technology
|
Page 2
www.rsisinternational.org
settings tends to center around the technical solutions like bias detection programs, explainable AI tools, and
compliance structures (Hermosilla et al., 2025; Saarela & Podgorelec, 2024). AI ethics in financial planning
remains fragmented; existing frameworks (e.g., OECD, IEEE) are general and lack sector-specific guidance.
While regulations like GDPR or Basel III offer compliance structures, they do not address the moral reasoning
needed for high-stakes decisions like retirement planning or risk profiling. Procedural tools (e.g., explainable AI,
fairness audits) support transparency but fall short of defining what ethically ought to be done. While significant,
these innovations fail to adequately confront the underlying ethics about what financial AI systems should do,
who should be held accountable, and what fairness should mean in financial decisions. No unified normative
ethics framework currently exists in this area that should inform AI's design, deployment, and governance in
financial planning settings. Morley et al. (2021) suggest that most AI ethics frameworks fail to provide domain-
specific guidance. Also, Jobin et al. (2019) reviewed 84 documents containing ethical principles or guidelines
for AI and found that no single, unified, or enforceable ethical framework exists across sectors.
This paper fills this crucial gap. It makes the case for a normative framework of ethics built around three key
issues: bias, transparency, and the use of human judgment. Rather than defining the issue technically or legally,
this paper borrows ideas from normative ethics, particularly theories of fairness, responsibility, and moral agency,
to make the case for an ethical approach to AI in financial planning. The objective goes beyond risk identification
to provide an ethically oriented direction for developers, financial firms, regulators, and other players in the
financial technology sector. The paper's contribution has two parts. Firstly, it presents a theoretically informed
ethical framework that incorporates fairness, transparency, and human agency into AI-informed financial
decisions. The framework has its roots in normative ethics and theory, but takes the form that might inform
practical use. Secondly, the paper considers how this framework might apply in real-world financial situations,
like automated investment recommendations or credit scoring.
Conceptual Foundations
AI in Financial Planning
Artificial Intelligence (AI) in financial planning refers to the use of computer programs that are capable of
conducting tasks that would normally involve human intelligence. Such tasks are pattern recognition, decision-
making, and forecasting (Najem et al., 2025). In financial services, AI programs are continuously incorporated
into services with the view of boosting precision, speed, and customization (Vuković et al., 2025). There are two
categories of AI applications in financial planning. The first is decision support systems. These systems assist
human financial advisors by providing advanced analytics and recommendations. They do not make the final
decision but offer data-driven insights. Examples include risk analytics tools and market trend predictors. The
other category includes decision automation systems. These are automated systems that make decisions without
direct human intervention. Robo-advisors are an obvious example. They apply algorithms to make investment
portfolio recommendations and adjustments, given user choices and market trends (Jia et al., 2022; Tao et al.,
2021). The other area where AI is employed includes credit scoring (Raji et al., 2024). The conventional credit
scoring approaches depend on fixed indicators such as repayment history and income. AI, however, employs
machine learning algorithms that deal with large, diverse, but relevantly disparate sets of information, including
social media and history of transactions. The systems are capable of arriving at more dynamic, inclusive scores
(Li et al., 2024). That, however, creates fairness issues, particularly when the information sources mirror past
discriminatory trends.
Portfolio management represents another key use. AI-driven robo-advisors like Betterment or Wealth front make
automated decisions about asset allocation with respect to market conditions and risk profiles (Lam, 2016).
Products like these make financial advice cheaper and more accessible. However, they also eliminate the human
factor in difficult financial choices, potentially impacting ways that compromise client trust and emotional
comfort (Ahmad et al., 2023). AI also finds use in risk management. Algorithms track volatility in the market,
flag potential instances of fraud, and evaluate systemic risk. Institutional financial use includes AI models
enabling stress scenarios for simulations, along with predicting economic trends. These, in turn, run typically in
real time, providing current information to decision-makers.
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XIII September 2025
Special Issue on Emerging Paradigms in Computer Science and Technology
|
Page 3
www.rsisinternational.org
However, when these instruments are utilized with non-transparent models or without regulation, they can
generate values that users cannot question or explain (Bahoo et al., 2024). It is useful at this point to differentiate
narrow AI from general AI. Today's financial programs are primarily narrow AI, which refers to a specific use
program. These are effective under defined constraints but do not possess general capabilities in giving advice,
in the form of general reasoning (Walton, 2018). This renders them effective, yet narrow, particularly where
unpredictable, new situations arise. AI in financial planning covers an enormous range of products and
capabilities. These range from aiding human advisors to complete automated processes. While possessing
advantages like effectiveness, accessibility, and speed in computing, they also pose an additional set of risks,
both practical and ethical. It is crucial, before determining its normative effects, to understand where AI becomes
incorporated into financial structures.
Ethics in Technology Use
Ethics in relation to technology, particularly Artificial Intelligence (AI), becomes more relevant as these
programs make decisions that have significant impacts on people's lives. The ability to assess the morality of
these decisions requires an education in ethical theory. Ethics involves the examination of right and wrong, and
how people should act. There are various branches within this area. Two important ones are normative ethics
and applied ethics (Chaddha & Agrawal, 2023). Normative ethics involves attempting to create theories and
principles, informing what people should do. It raises questions such as “What is the right thing to do here?” and
“What's our duty to others?” It does not become involved with specific situations but rather attempts to create
abstract rules and moral guidelines. For instance, it could investigate whether fairness should guide decisions or
whether decisions should maximize overall utility (Dempsey et al., 2023). By contrast, applied ethics takes these
theories and attempts to apply them in relation to specific, practical issues. Applied ethics is inherently concerned
with how AI systems impact human beings. It focuses on how the systems are developed, the processes, logic
of decision making, who makes decisions, how and to what extent. These range from controversies over data
privacy, algorithmic discrimination, or AI opacity (Bleher & Braun, 2023; Kazim & Koshiyama, 2021). In this
way, normative ethics lays the foundations, while applied ethics brings it into practice.
Several normative ethical theories are useful for understanding and critiquing AI systems in financial planning.
The three most relevant are deontology, utilitarianism, and virtue ethics. Each offers a different way to judge the
morality of AI development and use.
Deontology, as represented by scholars such as Immanuel Kant, stresses duty and rules. In deontological thought,
one fundamental notion holds that certain actions are morally obligatory or prohibited, regardless of the
consequences (Barrow & Khandhar, 2023). For instance, when an AI tool employed in the field of scoring arrives
at correct conclusions but discriminates against some persons, it remains unethical, according to the deontologist.
This is because it violates a moral duty to treat people equally and respect their rights. Deontology also
emphasizes transparency and accountability. Users and regulators should be able to know how decisions are
made. If an AI system cannot explain its decisions or allow individuals to contest them, it may be seen as morally
unacceptable under this view (Jedličková, 2024).
Utilitarianism, on the other hand, is focused on outcomes. It holds that the best action is the one that produces
the greatest good for the greatest number. When applied to AI in financial services, this theory looks at whether
an algorithm improves financial access, reduces costs, or benefits more people than it harms (Anshari et al.,
2022). A utilitarian might support a system that increases efficiency and reduces overall bias, even if a small
number of individuals are negatively affected. However, this approach can sometimes justify unfair treatment of
minorities if it benefits the majority. This tension raises ethical concerns in financial contexts, especially where
long-term inequalities may be reinforced by AI models trained on biased data (Card & Smith, 2020).
Virtue ethics takes a different approach. It does not focus on rules or outcomes but on the character of the people
and institutions involved. This theory asks whether the design, development, and use of AI reflect virtues like
responsibility, honesty, and integrity (Hagendorff, 2022). In the financial sector, it could be whether developers
are cautious when they are training models or whether financial firms are honest and transparent with users.
Virtue ethics encourages an ethics-aware culture, not technical compliance alone. It also prefers the approach of
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XIII September 2025
Special Issue on Emerging Paradigms in Computer Science and Technology
|
Page 4
www.rsisinternational.org
responsible innovation, where ethical thinking becomes part of innovation at the onset, not an afterthought
(Griffin et al., 2024).
Alongside traditional theories, there are also AI-specific ethical frameworks. These frameworks are designed to
address the unique features of AI systems, such as autonomy, opacity, and data dependency. One of the most
influential is the work of Luciano Floridi and colleagues, who developed principles like non-maleficence (do no
harm), beneficence, justice, and explicability (Floridi & Cowls, 2019). These principles combine elements of
normative ethics to guide the development of trustworthy AI.
Another milestone is the European Union’s AI Act, projecting risk-based regulation of AI applications. In April
2021, the European Commission proposed the first EU regulatory framework for AI, which was later publicized
as Regulation (EU) 2024/1689 (Regulation (EU) 2024) on 12 July 2024, establishing a governance structure and
setting out clear requirements for the Commission and the AI Office. It classifies AI systems in several risk
categories, ranging from minimal to unacceptable. The higher-risk ones, such as those used in employment
choices or credit ratings, face more stringent regulations. It aims at offering fairness, human oversight, and
explainability in automated decisions (Szadeczky & Bederna, 2025). The IEEE P7000 series is another
regulation. It was launched by the Institute of Electrical and Electronics Engineers (IEEE) as a set of standards
committed to embedding ethically relevant considerations into AI system building. For example, the IEEE P7001
standard touches on transparency, with the provision that users should be able to understand and question
decisions made by intelligent systems (Spiekermann, 2017). These guidelines often overlap with mainstream
ethical theories but are set down with technological deployments in focus.
These frameworks and theories lay a strong base for analyzing the ethics of AI in financial planning. They guide
not only what AI can do, but what it ought to do and under what circumstances. No one theory has all the answers,
but together, they permit an enriched, better-balanced ethical evaluation.
Key Ethical Tensions in Financial AI Systems
AI technologies in financial services hold several ethical tensions caused by the intersection between efficiency,
fairness, accountability, and autonomy. These tensions are not just technological difficulties but ethical
challenges that impact individual rights, institutional trust, as well as societal equity. Handling these issues
involves an understanding of the nature of AI systems and ways in which their use in financial settings creates
tensions that are ethically problematic.
Bias and Discrimination
Hanna et al. (2024) indicate that AI bias represents a key ethical issue. AI financial systems frequently use past
data to train algorithms. If such data includes past inequality or systematic discrimination, an AI system could
reproduce or even magnify bias. For example, the algorithms for scoring credit could act against minority
populations because of ingrained patterns in the training set. Even neutral-looking variables such as zip codes or
school names become proxies for race or socio-economic background, causing discriminatory decisions about
giving credit (Cristina et al., 2023). The dilemma lies between AI’s promise of efficiency and fairness demanded
in financial services. Companies may try to maximize prediction and outcomes, but this optimizes at the expense
of treating people fairly. Biases in algorithms could pass performance tests but consistently discriminate against
populations. The dilemma points to the limitation of exclusively data-based systems and the importance of
ethical regulation.
Transparency vs. Complexity
Most AI applied in finance, like deep learning models, act as “black boxes.” The inner mechanisms are not clear,
even to experts. This unpredictability causes an important ethical tension (Svetlova, 2022). Financial judgments
involve high stakes, like granting loans, distributing investments, and setting insurance rates. However, users
and affected persons frequently cannot see how to review the rationale behind decisions. This tension places
model precision and sophistication against the ethical need for explainability. Regulators and ethicists maintain
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XIII September 2025
Special Issue on Emerging Paradigms in Computer Science and Technology
|
Page 5
www.rsisinternational.org
that people should be entitled to comprehend decisions that impact their financial well-being. The EU's General
Data Protection Regulation (GDPR) has a "right to explanation," but it is difficult in practice when AI systems
are "black boxes" (Wachter et al., 2018). This also causes an issue with "information asymmetry." AI builders
and financial firms are normally better versed technically than their customers, causing uneven power dynamics.
Without clarity, users cannot provide informed consent or question decisions, eroding autonomy and
accountability.
Automation vs. Human Judgment
AI allows high levels of financial decision-making automation. Robo-advisors and automated loan processors
keep prices low and ensure 24/7 operations. There is an efficiency cost, however, in that human judgment is not
present (Maier et al., 2022). AI has no compassion, moral sense, or capacity to consider individual situations.
The outcome can be callous or rigid decisions. There is an ethical trade-off between speed/consistency and
contextual sensitivity/moral judgment (Farisco et al., 2020). For example, a human loan officer might waive an
unfavorable credit report because of strong individual circumstances. An AI program might lack that flexibility.
The tension is particularly relevant where financial distress intersects with sensitive populations. Also, AI
dependence has the potential to induce “automation bias,” where individuals follow algorithmic counsel even
when they suspect errors. It erodes the role of professional judgment in financial matters and obliterates
important checks and balances. The use of human-in-the-loop mechanisms has been proposed as a solution, but
it creates its own set of ethics around responsibility and liability (Salloch & Eriksen, 2024).
Data Privacy and Surveillance
Financial AI platforms deal with enormous quantities of personal information, such as income, consumption
patterns, credit history, and even social media behavior. While this information increases predictive ability, it
also creates questions about consent and privacy (Aldboush & Ferdous, 2023). The users do not know what
happens to their information, who accesses it, or how long it remains in storage. The ethical dilemma inherent
in this situation involves striking an equilibrium among data effectiveness in driving innovation, risk evaluation,
and individual rights over privacy, as well as data safeguard. Financial institutions can contend that data-led
personalization has advantages. However, without clear guidelines, information gathering becomes overbearing
and compulsory. AI-powered monitoring could result in profiling and manipulation, causing mistrust in financial
products. The dilemma gets compounded by governance gaps. While some places, such as the EU, possess
effective data safeguard regulations, others lack effective frameworks. Incompatibilities in governance over
information leave cross-border financial multinationals with an ethical dilemma, as they operate across borders
(Thein et al., 2024).
Accountability and Moral Responsibility
Machado et al. (2024) noted that one of the ethical conflicts in AI-based finance is accountability. When an
algorithm causes an adverse or discriminatory outcome, it is unclear what entity, if any, should be held
accountable—the programmer, the bank, the source of information, or the program. This diffusion compromises
redress with an ethical basis. In conventional financial frameworks, chains of accountability are clear. With AI,
responsibility becomes fragmented. This leads to what is sometimes called the “moral crumple zone,” where
human actors absorb blame for decisions made by opaque systems (Kaas, 2024). This tension threatens both
legal clarity and moral justice. There is growing consensus that ethical AI systems in finance must include
accountability frameworks—clear documentation, audit trails, and transparent governance structures. Without
these, financial AI remains a system where errors are difficult to trace and ethical violations are easy to deflect
(Cheong, 2024; Raji et al., 2020). These tensions are visually summarised in Figure 1, which illustrates the
ethical trade-offs that frequently emerge in the design and deployment of AI systems in financial planning.
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XIII September 2025
Special Issue on Emerging Paradigms in Computer Science and Technology
|
Page 6
www.rsisinternational.org
Figure 1: Key Ethical Tensions in Financial AI Systems
Core Ethical Issues in AI-Driven Financial Planning
Algorithmic Bias
Algorithmic bias refers to systematic and repeatable errors in AI outputs that unfairly disadvantage certain
groups (Ukanwa, 2024). In financial planning, such biases can influence important decisions like loan approvals,
credit scoring, investment risk profiling, and insurance pricing. While these systems are often marketed as
objective and data-driven, they can perpetuate or even amplify societal inequalities embedded in their training
data or decision rules (Cristina et al., 2023; Fuster et al., 2021). Bias can enter AI systems through several
pathways. One common source is the training data. If historical financial data reflects discriminatory lending
practices or underrepresentation of certain populations, an AI model trained on that data will likely learn and
reproduce those same patterns (de Castro Vieira et al., 2025; Nwafor et al., 2024). Feature selection also plays a
critical role. Even if sensitive variables like race or gender are removed, proxies such as ZIP codes, education
level, or employment history can indirectly encode discriminatory patterns (Wang et al., 2024). Additionally,
biased assumptions built into model architecture or optimization criteria (such as maximizing accuracy over
fairness) can further entrench inequality.
There are growing examples of algorithmic bias in real-world financial services. One well-known case involved
a major credit card company whose algorithm gave significantly lower credit limits to women than men, even
when both had similar financial profiles. In another case, a digital lending platform disproportionately rejected
minority applicants, despite their creditworthiness. Such disparities highlight how automated decision systems,
if left unchecked, can perpetuate racial and gender discrimination under the guise of efficiency and neutrality
(Cristina et al., 2023). From an ethical perspective, algorithmic bias raises concerns about justice, fairness, and
non-discrimination. Deontological ethics emphasizes respect for individuals’ rights and equal treatment under
rules. Under this view, an AI system that treats similar individuals differently based on irrelevant or prejudicial
factors violates the moral duty of fairness (Ebrahimi et al., 2024). Even if biased outcomes result in economic
efficiency (e.g., by targeting the most “profitable” borrowers), a utilitarian approach must weigh these benefits
against the broader social harms and loss of trust in financial institutions. The injustice suffered by individuals
unfairly denied access to financial opportunities cannot be justified by aggregate economic gain.
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XIII September 2025
Special Issue on Emerging Paradigms in Computer Science and Technology
|
Page 7
www.rsisinternational.org
Virtue ethics further emphasizes the moral character of decision-makers and institutions. Developers and
financial professionals who deploy AI systems bear a responsibility to build and use them with care, integrity,
and a commitment to social justice. Algorithmic fairness should not be an afterthought or optional feature; it is
a core requirement of responsible innovation. These encompass the use of fairness-aware machine learning,
auditing bias in systems, and engaging multi-stakeholders in the design of systems (Hagendorff, 2022). A
normative approach to this question holds that AI in financial planning should be transparent, non-discriminatory,
and fair. Financial institutions operate with immense power over people’s economic lives. When they use AI
tools, they take upon themselves the ethical responsibility of ensuring those tools do not deepen social
inequalities. Technical solutions such as de-biasing algorithms and explainable AI are important but insufficient.
Ethical supervision, human judgment, and accountability mechanisms should go along with technological
protection (Kowald et al., 2024).
Transparency and Explainability
Financial planning AI solutions frequently use sophisticated algorithms. These are often deep learning
algorithms that cannot be interpreted to explain how exactly a specific choice was made. These algorithms are
sometimes referred to as “black-box” algorithms because they cannot provide transparent explanations to both
financial experts and users (Černevičienė & Kabašinskas, 2024). In financial services, a lack of transparency
presents severe ethical challenges. The financial decisions influence people's access to credit, investment,
creditworthiness, as well as future financial stability. If an individual has been given a recommendation or choice
where the rationale cannot be explained, this compromises the issue of informed consent as well as erodes trust.
The ethical appeal for explainability in financial AI systems corresponds with higher values of autonomy and
accountability. From an autonomy standpoint, people should understand the rationale behind decisions made
about their financial lives. By knowing, they are in a better position to decide. Without explainability, autonomy
suffers. Moreover, a lack of explainability complicates the ability to catch errors or bias in the system, with
implications for accountability and fairness (Wachter et al., 2017). An example in practice includes the early
adoption of robo-advisors. Some customers complained about receiving portfolio recommendations that were
inconsistent with risk tolerance, with the system offering no adequate explanation (Boreiko & Massarotti, 2020).
In this scenario, the black box nature of the algorithm causes challenges in questioning the recommendation, not
knowing if the recommendation was correct, and worst, suffering financial loss. In ethics, this is considered
going beyond a technological issue. It becomes a morality issue. According to deontological ethics, people
should be respected and not considered passive recipients of decisions. Explainable AI allows fairness, respects
dignity, and supports the creation of accountable AI systems (D’Alessandro, 2024).
Human Judgment and Accountability
Growing dependence on AI in financial planning has also resulted in an increasing trend toward complete
automation. Algorithmic decision systems and robo-advisors are now able to administer portfolios, evaluate risk,
and provide investment product recommendations without human intervention (Boreiko & Massarotti, 2020).
While speed and efficiency are offered, questions about errors, bias, and loss of public trust arise when human
advisors' roles are eliminated. It is an important question whether the elimination of human supervision increases
financial systems' vulnerability to errors, bias, or loss of public trust. Elimination of human judgment has serious
risks. When errors are made, automated programs multiply small errors on a large scale. While humans are
sensitive to emotions, AI lacks sensitivity as well as the ability to interpret the distinctive circumstances of the
clients. It can result in technically correct but ethically inappropriate recommendations. The studies proved users
feel uncomfortable dealing with fully automated financial instruments, mainly when decisions lack clarity,
empathy, or understanding (Klingbeil et al., 2024; Zhu et al., 2024).
Ethical reasoning favors the continuation of human judgment. Empathy, understanding of context, and moral
accountability are human capabilities that cannot be fully emulated by machines. The question of moral agency
suggests that an individual has to be held responsible for decisions, particularly when causes of harm arise. It is
why proposals like human-in-the-loop and significant human control have been made (Santoni de Sio & Van
den Hoven, 2018). These frameworks keep humans within the decision-making continuum, with control over
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XIII September 2025
Special Issue on Emerging Paradigms in Computer Science and Technology
|
Page 8
www.rsisinternational.org
what happens and the right to override when called upon. Human judgment is needed not only to correct mistakes
but to maintain values like care, trust, and accountability in financial services.
Proposed Normative-Ethical Framework
Ethical Design Principles: The EFT Model
In response to the ethics issues in AI-based financial planning, this paper proposes a novel framework known as
the EFT Model. This is a normative-ethical framework constructed upon four pillars: Ethical Intent, Fairness,
Transparency, and Human Oversight. Each one aims at an essential area of ethics concern and has been crafted
to ensure the prudent advancement as well as use of financial AI systems.
Ethical Intent entails the incorporation of values into the design process right from the beginning. AI systems
should be constructed with the intention to advance social good, the welfare of the client, and professional ethics.
Developers, as well as financial institutions, should give active thought to the potential damage or abuse of their
systems. Ethical intent encompasses the values of beneficence and non-maleficence, commonly referenced in
bioethics, and is equally applicable in financial situations (Floridi et al., 2018; Jobin et al., 2019).
Fairness demands that AI systems be designed to prevent bias and discriminatory outcomes. This includes
regular audits of training data, validation of algorithms across demographic groups, and corrective actions where
disparities exist. Bias can enter through data, model selection, or even developer assumptions. Therefore, fairness
must be both a design goal and a regulatory requirement (Buolamwini & Gebru, 2018; Mehrabi et al., 2019).
Transparency is essential to build trust and enable accountability. Financial AI systems should disclose their
methods, criteria, and logic used in decision-making. Explainable AI (XAI) techniques are crucial in helping
clients understand why specific recommendations or rejections occur. Without transparency, it becomes
unfeasible to obtain informed consent, and users could disengage or resist the system (Madaan, 2025).
Human Oversight ensures that critical decisions do not occur in isolation from human judgment. Human-in-the-
loop systems retain a layer of interpretive control, especially for high-stakes decisions. Assigning responsibility
is part of this principle, helping ensure that when errors occur, accountability is traceable (Santoni de Sio & Van
den Hoven, 2018).
The EFT Model constitutes an operational manual for aligning AI instruments with financial services' ethical
values. The flow and organization of the suggested Ethical Framework (EFT) are diagrammatically represented
in Figure 2.
Figure 2: EFT Model for AI-Driven Financial Decision-Making
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XIII September 2025
Special Issue on Emerging Paradigms in Computer Science and Technology
|
Page 9
www.rsisinternational.org
Stakeholder Roles and Obligations
In an ethically sound financial AI landscape, every major player has an active role in designing secure, just, and
trustworthy systems. While their roles are distinct, they are interdependent and need to be congruent in order for
the system to operate optimally.
Developers are the first point of ethical duty. They create, write, and implement algorithms, with those decisions
affecting all users. It falls upon developers to make it fair, transparent, and accurate. That includes employing
varied sets of data, verifying bias, and applying explainable AI models (Heidari et al., 2019). In addition to
technological skills, developers should be educated in ethical thinking and innovation responsibility (Griffin et
al., 2024).
Ridzuan et al. (2024) noted that financial institutions are also promoters of AI tools. They are responsible for
integrating AI ethically into financial services. This means choosing vendors who comply with ethical standards,
conducting regular audits of AI performance, and ensuring that clients understand how recommendations are
made. Institutions must provide clear documentation and avenues for appeal when clients dispute results. They
also bear ultimate accountability for harm caused by the systems they use.
Regulators act as external guardians. Their role is to create policies that guide AI use, define fairness and
transparency standards, and enforce compliance. In fast-evolving fields like AI, regulators must also stay updated
and adapt rules to emerging risks. Regulatory sandboxes, for instance, allow testing AI systems under
supervision before full deployment (Yordanova & Bertels, 2023).
Clients also have responsibilities. As end-users, their decisions often rely on AI outputs. Therefore, digital
financial literacy is essential. Clients must understand basic concepts like risk profiling, data sharing, and AI
limitations. Institutions must support this by offering user-friendly tools and educational resources (Amnas et
al., 2024).
Shared ethical responsibility ensures that no single group bears the burden alone. A multi-actor approach
strengthens trust and accountability across the financial AI ecosystem.
Ethical Decision-Making Flow
Financial decisions can have significant consequences, especially when involving AI. A systematic process for
ethical decisions reduces damage, contains risk, and preserves human control. The three-stage flowchart consists
of Trigger Points, Ethical Checkpoints, and Escalation Paths.
The first level is the trigger point, where an AI financial system has encountered a high-stakes or ethically
sensitive situation. These could be investment choices beyond some threshold, loan grants, or retirement account
disbursements. These situations flag the system automatically for an extra level of check.
The second phase includes ethical checkpoints. These are integrated criteria that assess the recommendation's
fairness level, transparency, and accuracy. For example, to what extent has the algorithm explained its
recommendation adequately? Are there signs of bias, conflicting information? These checkpoints are like
internal auditing.
In the event that a checkpoint fails, the process escalates, calling upon human review. It then becomes the
responsibility of a financial advisor or compliance officer to intervene, reviewing the system manually. Human
reviewers are trained in applying contextual judgment, understanding client worries, and making ethically
sensitive choices. This phase maintains significant human control with the prevention of loss of moral agency in
automation (Santoni de Sio & Van den Hoven, 2018).
This flow allows ethical vigilance without sacrificing efficiency. It can be presented in a simple decision tree
diagram. The visual shows where automation operates, when ethical checks apply, and when human input is
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XIII September 2025
Special Issue on Emerging Paradigms in Computer Science and Technology
|
Page 10
www.rsisinternational.org
required. Institutions can adapt this structure to fit their service types and risk levels. The flowchart in Figure 3
illustrates the ethical decision-making process embedded within financial AI systems, highlighting key
checkpoints, escalation triggers, and pathways for ensuring accountability and transparency in automated high-
stakes decisions.
Figure 3: Ethical Decision-Making Flow for Financial AI Systems
Application of Framework to Real-World Scenarios
The EFT model can be applied to the following real-world scenarios to address the issues created by the use of
AI in financial services.
Robo-advisor investment portfolio with opaque methodology
Many retail investors use robo-advisors to manage their investments. However, some of these systems do not
explain how portfolio recommendations are generated. This lack of transparency can cause confusion and
distrust, especially during market downturns. Applying the EFT framework, the Transparency pillar would
require explainable AI (XAI) techniques, where the user is shown a clear rationale behind each investment choice.
Ethical responsibility dictates that the developer include audit logs and visual summaries. The Fairness pillar
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XIII September 2025
Special Issue on Emerging Paradigms in Computer Science and Technology
|
Page 11
www.rsisinternational.org
ensures that all users receive equitable risk profiles regardless of financial literacy. This level of openness can
reduce panic-driven withdrawals and build user confidence.
AI loan algorithm biased against minority applicants
In 2019, Apple Card faced allegations that women were given lower credit limits than men with similar financial
profiles (BBC, 2019). Similar concerns have been raised in U.S. mortgage lending data (Liu & Liang, 2025).
Using the EFT framework, the Fairness component mandates proactive algorithmic audits and bias detection
protocols. Transparency would require lenders to disclose how creditworthiness is calculated, especially when
traditional credit scores are supplemented with alternative data. Human oversight becomes critical in high-impact
decisions, such as loan approvals. It would allow flagged decisions to be reviewed manually. This helps restore
public trust, as well as comply with regulatory requirements.
Automated retirement recommendations without individual context
Some financial planning websites supply retirement horizons without considering factors that may be specific
to certain users, like continued illness, dependents, or late-career turbulence (Gorry & Leganza, 2024). These
present ethical concerns with individualization and human dignity. Using the EFT paradigm, Ethical
responsibility may involve requiring users to provide contextual variables. Fairness means requiring the system
to account for life-stage variation. This keeps it relevant and prevents damage caused by one-size financial advice.
Implications
The Ethical Framework (EFT) presents practical, regulatory, and research avenues for increasing prudent AI
application in financial services.
For Practice
Fintech developers are primarily responsible for implementing the EFT framework at the design level. Ethical
values should be infused in system architecture. That involves utilizing varied training data, fairness testing, and
explainable model use. The developers should ensure that decision rules are explainable and outcomes are
interpretable. UX interfaces should also exhibit transparency with explanations and opt-out options provided for
users. These measures avoid eroding trust and facilitate responsible innovation.
Financial planners should not depend exclusively on automated systems. Rather, active supervision should be
preserved, particularly in high-risk decisions. The human-in-the-loop concept guarantees that clients should be
able to question automated recommendations. The use of the EFT framework should help planners decide when
technology should be employed and when individual judgment should be exercised. It keeps the interests of the
client secure while ensuring the duty of care. System output should also be frequently reviewed to identify
potential causes of harm.
For Policy and Regulation
The EFT framework aligns closely with Article 22 of the GDPR, providing individuals with the right not to be
made subject to decisions under fully automated processes. Even the EU AI Act preserves financial decision-
making tools as having the potential for high risk and specifies the need for transparency, fairness, and human
oversight. In America, the FTC and Consumer Financial Protection Bureau (CFPB) guidelines provide similar
warnings about the black box mechanisms of AI leading to discrimination. The EFT model offers usable tools
to meet such legal mandates. It could also complete ISO/IEC 42001 AI management systems standards with
specific ethical directions, along with procedures for safeguards.
For Research
There is a need for empirical research to confirm the EFT framework in multiple financial settings. Future
research should experiment with how the framework shapes user trust, decision making, and bias minimization.
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XIII September 2025
Special Issue on Emerging Paradigms in Computer Science and Technology
|
Page 12
www.rsisinternational.org
It is possible to combine the use of EFT metrics with AI auditing instruments like IBM’s AI Fairness 360 or
Google’s What-If Tool. By measuring ethical indicators, researchers can create guidelines for responsible
Fintech. It will also help contribute to the development of industry-wide certification programs.
CONCLUSION
Rising AI in financial services requires an unambiguous and consistent ethics strategy. As automation spreads,
the dangers of bias, lack of transparency, and loss of human agency also escalate. There is a need for a normative
ethical framework, ensuring that AI systems are compatible with fundamental human values and regulations.
The proposed Ethical Framework (EFT), grounded in fairness, transparency, and human supervision, directly
confronts the issues. It encourages algorithmic review, informs system logic, and enforces human accountability
in decisions. These values are not only moral obligations but also practical instruments for constructing user
trust and institutional legitimacy. Fair and transparent AI underpins regulation compliance while enabling clients
to make independent decisions. Human agency remains fundamental, where empathy and moral judgment are
important, especially in high-risk situations. This novel framework fills an important gap in current practice and
suggests a direction towards accountable AI governance in financial sectors.
REFERENCES
1. Agarwal, V. (2024). Fair or Flawed? Assessing AI’s Impact on Credit Decisions. International Journal of
Computer Trends and Technology, 72(Dec 12, 2024), 128–132. https://doi.org/10.14445/22312803/IJCTT-
V72I12P115
2. Ahmad, S. F., Han, H., Alam, M. M., Rehmat, Mohd. K., Irshad, M., Arraño-Muñoz, M., & Ariza-Montes,
A. (2023). Impact of Artificial Intelligence on Human Loss in Decision making, Laziness and Safety in
Education. Humanities and Social Sciences Communications, 10(1), 1–14. https://doi.org/10.1057/s41599-
023-01787-8
3. Aldboush, H. H. H., & Ferdous, M. (2023). Building Trust in Fintech: An Analysis of Ethical and Privacy
Considerations in the Intersection of Big Data, AI, and Customer Trust. International Journal of Financial
Studies, 11(3). https://doi.org/10.3390/ijfs11030090
4. Amnas, M. B., Selvam, M., & Parayitam, S. (2024). FinTech and Financial Inclusion: Exploring the
Mediating Role of Digital Financial Literacy and the Moderating Influence of Perceived Regulatory Support.
Journal of Risk and Financial Management, 17(3), 108. https://doi.org/10.3390/jrfm17030108
5. Anshari, M., Hamdan, M., Ahmad, N., Ali, E., & Haidi, H. (2022). COVID-19, Artificial intelligence, Ethical
Challenges and Policy Implications. AI & Society, 38(2). https://doi.org/10.1007/s00146-022-01471-6
6. Bahoo, S., Cucculelli, M., Goga, X., & Mondolo, J. (2024). Artificial intelligence in Finance: a
comprehensive review through bibliometric and content analysis. Artificial Intelligence in Finance: A
Comprehensive Review through Bibliometric and Content Analysis, 4(2). https://doi.org/10.1007/s43546-
023-00618-x
7. Barrow, J. M., & Khandhar, P. B. (2023, August 8). Deontology. National Library of Medicine; StatPearls
Publishing. https://www.ncbi.nlm.nih.gov/books/NBK459296/
8. BBC. (2019, November 11). Apple’s “sexist” credit card probed by regulator. BBC News.
https://www.bbc.com/news/business-50365609
9. Bleher, H., & Braun, M. (2023). Reflections on Putting AI Ethics into Practice: How Three AI Ethics
Approaches Conceptualize Theory and Practice. Science and Engineering Ethics, 29(3).
https://doi.org/10.1007/s11948-023-00443-3
10. Boreiko, D., & Massarotti, F. (2020). How Risk Profiles of Investors Affect Robo-Advised Portfolios.
Frontiers in Artificial Intelligence, 3. https://doi.org/10.3389/frai.2020.00060
11. Buolamwini, J., & Gebru, T. (2018). Gender Shades: Intersectional Accuracy Disparities in Commercial
Gender Classification. Proceedings of Machine Learning Research, 81(1), 1–15.
https://proceedings.mlr.press/v81/buolamwini18a/buolamwini18a.pdf
12. Card, D., & Smith, N. A. (2020). On Consequentialism and Fairness. Frontiers in Artificial Intelligence, 3(1),
1–11. https://doi.org/10.3389/frai.2020.00034
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XIII September 2025
Special Issue on Emerging Paradigms in Computer Science and Technology
|
Page 13
www.rsisinternational.org
13. Černevičienė, J., & Kabašinskas, A. (2024). Explainable artificial intelligence (XAI) in finance: a systematic
literature review. Artificial Intelligence Review, 57(8). https://doi.org/10.1007/s10462-024-10854-8
14. Chaddha, R., & Agrawal, G. K. (2023). Ethics and morality. Indian Journal of Orthopaedics, 57(11), 1707–
1713. https://doi.org/10.1007/s43465-023-01004-3
15. Cheong, B. C. (2024). Transparency and accountability in AI systems: safeguarding wellbeing in the age of
algorithmic decision-making. Frontiers in Human Dynamics, 6. https://doi.org/10.3389/fhumd.2024.1421273
16. Cristina, A., Gomes, M., & Rigobon, R. (2023). Algorithmic discrimination in the credit domain: what do we
know about it? AI & Society, 39, 2059–2098. https://doi.org/10.1007/s00146-023-01676-3
17. D’Alessandro, W. (2024). Deontology and safe artificial intelligence. Philosophical Studies.
https://doi.org/10.1007/s11098-024-02174-y
18. de Castro Vieira, J. R., Barboza, F., Cajueiro, D., & Kimura, H. (2025). Towards Fair AI: Mitigating Bias in
Credit Decisions—A Systematic Literature Review. Journal of Risk and Financial Management, 18(5), 228.
https://doi.org/10.3390/jrfm18050228
19. Dempsey, R. P., Eskander, E. E., & Dubljević, V. (2023). Ethical decision-making in law enforcement: A
scoping review. Psych, 5(2), 576–601. https://doi.org/10.3390/psych5020037
20. Ebrahimi, S., Abdelhalim, E., Hassanein, K., & Head, M. (2024). Reducing the incidence of biased
algorithmic decisions through feature importance transparency: an empirical study. European Journal of
Information Systems, 34(4), 636–664. https://doi.org/10.1080/0960085x.2024.2395531
21. Farisco, M., Evers, K., & Salles, A. (2020). Towards Establishing Criteria for the Ethical Analysis of Artificial
Intelligence. Science and Engineering Ethics, 26(5), 2413–2425. https://doi.org/10.1007/s11948-020-00238-
w
22. Floridi, L., & Cowls, J. (2019). A unified framework of five principles for AI in society. Harvard Data Science
Review, 1(1). https://doi.org/10.1162/99608f92.8cd550d1
23. Floridi, L., Cowls, J., Beltrametti, M., Chatila, R., Chazerand, P., Dignum, V., Luetge, C., Madelin, R.,
Pagallo, U., Rossi, F., Schafer, B., Valcke, P., & Vayena, E. (2018). AI4People—An Ethical Framework for
a Good AI Society: Opportunities, Risks, Principles, and Recommendations. Minds and Machines, 28(4),
689–707. https://doi.org/10.1007/s11023-018-9482-5
24. Fuster, A., Goldsmith-Pinkham, P., Ramadorai, T., & Walther, A. (2021). Predictably Unequal? The Effects
of Machine Learning on Credit Markets. The Journal of Finance, 77(1), 5–47.
https://doi.org/10.1111/jofi.13090
25. Giarmoleo, F. V., Ferrero, I., Rocchi, M., & Pellegrini, M. (2024). What ethics can say on artificial
intelligence: Insights from a systematic literature review. Business and Society Review, 129(2), 258–292.
https://doi.org/10.1111/basr.12336
26. Gladstone, J., & Hundtofte, C. S. (2023). A lack of financial planning predicts increased mortality risk:
Evidence from cohort studies in the United Kingdom and United States. PLoS ONE, 18(9).
https://doi.org/10.1371/journal.pone.0290506
27. Gorry, A., & Leganza, J. M. (2024). How do life events affect retirement timing? Evidence from expectations
data. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.4815140
28. Griffin, T. A., Green, B. P., & Welie, J. V. M. (2024). The ethical wisdom of AI developers. AI and Ethics,
5, 1087–1097. https://doi.org/10.1007/s43681-024-00458-x
29. Hagendorff, T. (2022). A Virtue-Based Framework to Support Putting AI Ethics into Practice. Philosophy &
Technology, 35(3). https://doi.org/10.1007/s13347-022-00553-z
30. Hanna, M., Pantanowitz, L., Jackson, B., Palmer, O., Visweswaran, S., Pantanowitz, J., Deebajah, M., &
Rashidi, H. (2024). Ethical and Bias Considerations in Artificial intelligence/machine Learning. Modern
Pathology, 38(3), 1–13. ScienceDirect. https://doi.org/10.1016/j.modpat.2024.100686
31. Heidari, H., Loi, M., Gummadi, K. P., & Krause, A. (2019). A Moral Framework for Understanding Fair ML
through Economic Models of Equality of Opportunity. Proceedings of the Conference on Fairness,
Accountability, and Transparency, 181–190. https://doi.org/10.1145/3287560.3287584
32. Hermosilla, P., Díaz, M., Berríos, S., & Allende-Cid, H. (2025). Use of Explainable Artificial Intelligence for
Analyzing and Explaining Intrusion Detection Systems. Computers, 14(5).
https://doi.org/10.3390/computers14050160
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XIII September 2025
Special Issue on Emerging Paradigms in Computer Science and Technology
|
Page 14
www.rsisinternational.org
33. Jedličková, A. (2024). Ethical approaches in designing autonomous and intelligent systems: a comprehensive
survey towards responsible development. AI & Society, 40, 2703–2716. https://doi.org/10.1007/s00146-024-
02040-9
34. Jia, T., Wang, C., Tian, Z., Wang, B., & Tian, F. (2022). Design of Digital and Intelligent Financial Decision
Support System Based on Artificial Intelligence. Computational Intelligence and Neuroscience, 2022, 1–7.
https://doi.org/10.1155/2022/1962937
35. Jobin, A., Ienca, M., & Vayena, E. (2019). The Global Landscape of AI Ethics Guidelines. Nature Machine
Intelligence, 1(9), 389–399. https://doi.org/10.1038/s42256-019-0088-2
36. Kaas, M. H. L. (2024). The perfect technological storm: artificial intelligence and moral complacency. Ethics
and Information Technology, 26(3). https://doi.org/10.1007/s10676-024-09788-0
37. Kazim, E., & Koshiyama, A. S. (2021). A High-level Overview of AI Ethics. Patterns, 2(9), 1–12.
https://doi.org/10.1016/j.patter.2021.100314
38. Klingbeil, A., Grützner, C., & Schreck, P. (2024). Trust and Reliance on AI — an Experimental Study on the
Extent and Costs of Overreliance on AI. Computers in Human Behavior, 160, 108352.
https://doi.org/10.1016/j.chb.2024.108352
39. Kowald, D., Scher, S., Pammer-Schindler, V., Müllner, P., Waxnegger, K., Demelius, L., Fessl, A., Toller,
M., Mendoza, G., Šimić, I., Sabol, V., Trügler, A., Veas, E., Kern, R., Nad, T., & Kopeinik, S. (2024).
Establishing and evaluating trustworthy AI: overview and research challenges. Frontiers in Big Data, 7.
https://doi.org/10.3389/fdata.2024.1467222
40. Lam, J. W. (2016, April 4). Robo-Advisors: A Portfolio Management Perspective. Yale Department of
Economics. https://economics.yale.edu/sites/default/files/2023-
01/Jonathan_Lam_Senior%20Essay%20Revised.pdf
41. Li, C., Wang, H., Jiang, S., & Gu, B. (2024). The Effect of AI-Enabled Credit Scoring on Financial Inclusion:
Evidence from an Underserved Population of over One Million. AIS Electronic Library.
https://aisel.aisnet.org/misq/vol48/iss4/25/
42. Liu, Z., & Liang, H. (2025). Are credit scores gender-neutral? Evidence of mis-calibration from alternative
and traditional borrowing data. Journal of Behavioral and Experimental Finance, 47, 101081.
https://doi.org/10.1016/j.jbef.2025.101081
43. Machado, J., Sousa, R., Peixoto, H., & António Abelha. (2024). Ethical Decision-Making in Artificial
Intelligence: A Logic Programming Approach. AI, 5(4), 2707–2724. https://doi.org/10.3390/ai5040130
44. Madaan, H. (2025, February 14). XAI: Bringing Transparency And Trust To Algorithmic Decisions. Forbes.
https://www.forbes.com/councils/forbestechcouncil/2025/02/14/the-rise-of-explainable-ai-bringing-
transparency-and-trust-to-algorithmic-decisions/
45. Maier, T., Menold, J., & McComb, C. (2022). The Relationship Between Performance and Trust in AI in E-
Finance. Frontiers in Artificial Intelligence, 5. https://doi.org/10.3389/frai.2022.891529
46. Mehrabi, N., Morstatter, F., Saxena, N., Lerman, K., & Galstyan, A. (2019). A Survey on Bias and Fairness
in Machine Learning. ArXiv:1908.09635 [Cs]. https://arxiv.org/abs/1908.09635
47. Morley, J., Floridi, L., Kinsey, L., & Elhalal, A. (2021). From what to how: An overview of AI ethics tools,
methods and research to translate principles into practices. AI and Society, 36, 59–72.
48. Najem, R., Bahnasse, A., Amr, M. F., & Talea, M. (2025). Advanced AI and big data techniques in E-finance:
a comprehensive survey. Discover Artificial Intelligence, 5(1). https://doi.org/10.1007/s44163-025-00365-y
49. Nallakaruppan, M. K., Chaturvedi, H., Grover, V., Balusamy, B., Jaraut, P., Bahadur, J., Meena, V. P., &
Hameed, I. A. (2024). Credit Risk Assessment and Financial Decision Support Using Explainable Artificial
Intelligence. Risks, 12(10). https://doi.org/10.3390/risks12100164
50. Nwafor, C. N., Nwafor, O., & Brahma, S. (2024). Enhancing transparency and fairness in automated credit
decisions: an explainable novel hybrid machine learning approach. Scientific Reports, 14(1).
https://doi.org/10.1038/s41598-024-75026-8
51. Raji, A. A. H., Alabdoon , A. H. F., & Almagtome, A. (2024). AI in Credit Scoring and Risk Assessment:
Enhancing Lending Practices and Financial Inclusion. International Conference on Knowledge Engineering
and Communication Systems, 15, 1–7. https://doi.org/10.1109/ickecs61492.2024.10616493
52. Raji, I. D., Smart, A., White, R. N., Mitchell, M., Gebru, T., Hutchinson, B., Smith-Loud, J., Theron, D., &
Barnes, P. (2020). Closing the AI Accountability Gap: Defining an End-to-End Framework for Internal
Algorithmic Auditing. ArXiv:2001.00973 [Cs]. https://doi.org/10.48550/arXiv.2001.00973
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue XIII September 2025
Special Issue on Emerging Paradigms in Computer Science and Technology
|
Page 15
www.rsisinternational.org
53. Ridzuan, N. N., Masri, M., Anshari, M., Fitriyani, N. L., & Syafrudin, M. (2024). AI in the Financial Sector:
The Line between Innovation, Regulation and Ethical Responsibility. Information, 15(8), 432.
https://doi.org/10.3390/info15080432
54. Saarela, M., & Podgorelec, V. (2024). Recent Applications of Explainable AI (XAI): A Systematic Literature
Review. Applied Sciences, 14(19). https://doi.org/10.3390/app14198884
55. Sahu, M. K. (2024). AI-Based Robo-Advisors: Transforming Wealth Management and Investment Advisory
Services. Journal of AI-Assisted Scientific Discovery, 4(1), 379–411.
https://www.scienceacadpress.com/index.php/jaasd/article/view/142
56. Salloch, S., & Eriksen, A. (2024). What Are Humans Doing in the Loop? Co-Reasoning and Practical
Judgment When Using Machine Learning-Driven Decision Aids. American Journal of Bioethics, 24(9), 1–
12. https://doi.org/10.1080/15265161.2024.2353800
57. Santoni de Sio, F., & van den Hoven, J. (2018). Meaningful Human Control over Autonomous Systems: A
Philosophical Account. Frontiers in Robotics and AI, 5(15). https://doi.org/10.3389/frobt.2018.00015
58. Spiekermann, S. (2017). IEEE P7000—The First Global Standard Process for Addressing Ethical Concerns
in System Design. Proceedings, 1(3), 159. https://doi.org/10.3390/is4si-2017-04084
59. Svetlova, E. (2022). AI ethics and systemic risks in finance. AI and Ethics, 2(4), 713–725.
https://doi.org/10.1007/s43681-021-00129-1
60. Szadeczky, T., & Bederna, Z. (2025). Risk, regulation, and governance: evaluating artificial intelligence
across diverse application scenarios. Security Journal, 38(1). https://doi.org/10.1057/s41284-025-00495-z
61. Tao, R., Su, C.-W., Xiao, Y., Dai, K., & Khalid, F. (2021). Robo advisors, algorithmic trading and investment
management: Wonders of fourth industrial revolution in financial markets. Technological Forecasting and
Social Change, 163, 120421. https://doi.org/10.1016/j.techfore.2020.120421
62. Thein, H. H., Grosman, A., Sosnovskikh, S., & Klarin, A. (2024). Should we stay or should we exit?
Dilemmas faced by multinationals under sanctioned regimes. Journal of World Business, 59(6), 101585.
https://doi.org/10.1016/j.jwb.2024.101585
63. Ukanwa, K. (2024). Algorithmic Bias: Social Science Research Integration Through The 3-D Dependable AI
Framework. Current Opinion in Psychology, 58, 101836. https://doi.org/10.1016/j.copsyc.2024.101836
64. Vuković, D. B., Dekpo-Adza, S., & Matović, S. (2025). AI integration in financial services: a systematic
review of trends and regulatory challenges. Humanities and Social Sciences Communications, 12(1).
https://doi.org/10.1057/s41599-025-04850-8
65. Wachter, S., Mittelstadt, B., & Floridi, L. (2017). Why a Right to Explanation of Automated Decision-Making
Does Not Exist in the General Data Protection Regulation. International Data Privacy Law, 7(2), 76–99.
https://doi.org/10.1093/idpl/ipx005
66. Wachter, S., Mittelstadt, B., & Russell, C. (2018). Counterfactual Explanations without Opening the Black
Box: Automated Decisions and the GDPR. Harvard Journal of Law & Technology.
https://arxiv.org/abs/1711.00399
67. Walton, P. (2018). Artificial Intelligence and the Limitations of Information. Information, 9(12), 332.
https://doi.org/10.3390/info9120332
68. Wang, X., Wu, Y. C., Ji, X., & Fu, H. (2024). Algorithmic discrimination: examining its types and regulatory
measures with emphasis on US legal practices. Frontiers in Artificial Intelligence, 7, 1320277.
https://doi.org/10.3389/frai.2024.1320277
69. Yordanova, K., & Bertels, N. (2023). Regulating AI: Challenges and the Way Forward Through Regulatory
Sandboxes. Law, Governance and Technology Series, 58, 441–456. https://doi.org/10.1007/978-3-031-
41264-6_23
70. Zhu, H., Vigren, O., & Söderberg, I.-L. (2024). Implementing artificial intelligence empowered financial
advisory services: A literature review and critical research agenda. Journal of Business Research, 174.
https://doi.org/10.1016/j.jbusres.2023.114494
