See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/378488023

Exploring AI and Machine Learning Applications in Banking: A Comprehensive

Review of Literature

Article in International Journal of Advanced Scientific Research & Development (IJASRD) · February 2024
DOI: 10.5281/zenodo.10707332

CITATIONS READS
5 1,316

2 authors:

Tayyab Muhammad Stephanie Ness

Southern Methodist University Diplomatic Academy of Vienna
13 PUBLICATIONS 325 CITATIONS 31 PUBLICATIONS 505 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Tayyab Muhammad on 26 February 2024.

The user has requested enhancement of the downloaded file.

Volume 9, Issue 2, February – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165

Exploring AI and Machine Learning Applications in

Banking: A Comprehensive Review of Literature
Author#1 Author#2
Tayyab Muhammad Stephanie Ness
Department of Electrical and Computer Engineering Diplomatische Akademie
Southern Methodist University, Dallas, US
https://orcid.org/0009-0003-1753-3241

Author#3 Author#4
Mykola Volkivskyi Yulu Gong
Taras Shevchenko National University of Kyiv Computer & Information Technology
Northern Arizona University

Abstract:- The use of Artificial Intelligence (AI) and make effective decisions. ML and AI are poised to
Machine Learning (ML) within the Banking sector has revolutionize the banking industry with impressive outcomes.
dramatically transformed the industry in recent times. AI This article will reveal the ways in which Machine Learning
has catalyzed new advancements on both the back and and Artificial Intelligence are used in various areas of the
front ends of financial operations. This article seeks to banking industry and will discuss how these institutions utilize
explore the significant impact of artificial intelligence computational intelligence to enhance their business
within the banking industry. It will look at how these operations (Castelli et al., 2016; Donepudi, 2016).
abilities further drive efficiency, allowing professionals in Undoubtedly, ML and AI have brought about a major
the industry to enhance the experience for their customer transformation in the banking sector. This revolution has
base. It will explore different use cases within the space significantly enhanced the banking experience in numerous
and paint a picture for the future as computational ways. While many financial institutions are still struggling to
intelligence continues to hold a key role in decision-making embrace the use of computational intelligence technologies, its
where accuracy is increasingly important. The banking implementation is rapidly spreading throughout the industry.
industry has traditionally been an area conservative about
the implementation of new technology due to security Hence, it is justified to state that the financial and
implications. However, this has begun to change as the banking services have experienced a major transformation due
banking industry begins to consolidate these technologies to the implementation of Machine Learning (ML) and
can provide an advantage in the market and catch the eye Artificial Intelligence (AI). Significantly, the expansion of
of a public who is using traditional banks less and opting fintech companies is a major factor in the ongoing
to reside with tech and fin tech companies. Understanding transformation that we are witnessing. For example, a recent
what AI and ML are is key to getting what these Q2 analysis of India's fintech sector showed a significant
technologies are about. This paper provides multiple real- increase in investment with 32 transactions completed during
world examples of how machine learning is likely to upend the same time frame, indicating strong growth in the financial
banking. By ingesting files and recognizing patterns, ML services industry. Traditional banks are rapidly adopting
algorithms help banks predict behavior and make choices technology such as Chatbot, but fintech companies, which
that will help redefine what banking is in the modern age. have long embraced AI, are playing an important role in
driving innovation and contributing significantly to financial
I. INTRODUCTION intelligence (Das et al., 2015).

In recent years, there has been significant growth in the Firstly, it is essential to understand the concepts of AI
use of Machine Learning (ML) and Artificial Intelligence (AI) and ML. John McCarthy, who is considered the father of AI,
in the financial industry. These institutions have leveraged describes it as the science and engineering involved in creating
their significant power to provide business solutions for both intelligent machines, particularly intelligent computer
front-end and back-end processes, ultimately enhancing programs (Kumar & Chandrakala, 2016). In general, AI
efficiency and elevating the customer experience. Lately, we involves transforming a computer into a robot or using
have seen how computational intelligence has become the key programming codes to make computers think and act with
factor in gaining a competitive advantage through its ability to human-like intelligence. AI involves creating computer

IJISRT24FEB1225 www.ijisrt.com 1057

 Volume 9, Issue 2, February – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
software and systems that can replicate human behavior and  AML and Fraud Pattern Detection
problem-solving abilities by studying human thinking patterns, AML is a collection of measures and rules created to
learning methods, and cognitive capabilities. In essence, prevent the laundering of money obtained through unlawful
Artificial Intelligence is enabling computers to develop both activities. Many banks use AI-powered mechanisms, which
intellectual and emotional capabilities. While AI and ML are are more resilient and sophisticated than traditional AML
often used as synonyms, they actually represent different models. With ongoing advancements and enhancements in AI,
concepts. these systems are poised to become more precise and efficient.
A large number of countries have started to adopt the potential
A common belief is that the basis of machine learning is of AI and ML in identifying and preventing fraud. The
the idea that machines have the ability to learn. Historically, National Stock Exchange of India has recently revealed plans
machine learning and Artificial Intelligence have been in to implement policies that will allow it to use machine
existence for over fifty years, as indicated by their learning to detect market trends and oversee the exchange to
fundamental principles. In 1959, IBM scientist Arthur Samuel prevent manipulation of its High-Frequency Trading (HFT)
was the first to utilize Machine Learning by creating a solution markets. They are enhancing their security by integrating
for a checkers game and publishing it (Arthur Samuel, 1959). artificial intelligence and machine learning into their
He described how, for the first time, a computer was able to surveillance system.
play checkers against humans and emerge victorious. Over
time, programmers have developed increasingly advanced  Personalized Banking and Automation
systems that allow machines to perform tasks typically done This could be an area where AI has excelled with its
by humans. Another well-known example is the ancient board creative solutions and strategies for providing banking users
game "Go," which has existed for over 2500 years. It is more with more accessible and convenient options. AI is making a
complex and strategic than chess, leading to the belief that no significant impact on the banking industry by offering tailored
computer could surpass a human in the game of "Go. " Four services to customers, such as chatbots that offer self-service
years ago, the long-held belief was shattered when a computer solutions, ultimately decreasing the burden on call centers.
program called AlphaGo defeated an 18-time world champion Nowadays, voice-activated virtual assistants are offering more
in a decisive 4 to 1 victory (Taher-Uz-Zaman et al., 2014). intelligent support to customers in all their interactions.
Machine learning has sped up the transformation of banking Whether it's verifying balances, setting up payments,
operations. Machine learning in this area uses past data and reviewing account activity, and other related tasks.
behaviors to forecast trends and assist in making decisions Furthermore, in today's market, there are numerous
(SEC Speech, 2016). applications available that provide customized financial
guidance. An AI application has been developed that monitors
II. ROLE OF ARTIFICIAL INTELLIGENCE AND individual earnings, monthly costs, and spending behavior,
MACHINE LEARNING IN BANKING providing tailored recommendations and financial advice for
an optimized financial strategy. It's not surprising that industry
As AI and ML are increasingly adopted by financial leaders are now incorporating robotic process automation into
institutions, the world is witnessing a growing range of their future strategies to cut operational expenses and improve
applications for these technologies daily. Nevertheless, the productivity through intelligent character recognition. The
potential risks associated with both are on the rise. integration of AI and ML in the banking industry reduces the
Computational intelligence technology is rapidly expanding its risks of human error in tasks that require frequent repetition.
influence beyond the banking sector and into industries like
insurance and capital markets. Artificial intelligence and  Customer Recommendations
machine learning are widely used in the banking sector to The use of computational intelligence has significantly
automate processes, conduct analysis, and facilitate decision- contributed to the development of recommendation engines in
making, ultimately leading to the development of innovative the banking industry. The process involves using historical
business strategies. Purdy & Daugherty (2016) found that user data to generate personalized recommendations that align
artificial intelligence and machine learning will shape the with the user's preferences and past interactions.
future of customer interactions in banking. A recent study Recommendation engines have long been a key factor in the
conducted by the BCG consulting group indicates that China increased profits that numerous banks have experienced over
is significantly leading the way in the implementation of AI the years.
and ML in its financial industry, particularly in the area of
fintech, compared to other countries. The study indicates that  Risk Management
by 2027, approximately a quarter of the finance job market The fundamental principles of AI and ML involve
will have evolved due to the significant impact of AI and ML learning from past data. It is no surprise that ML and AI are
in improving efficiency and automating processes. Below are rapidly taking over the banking industry, as they excel in
several uses of Artificial Intelligence (AI) and Machine managing financial records and transactions. An example
Learning (ML) in the banking industry (Yu et al., 2016) would be how credit cards work. For a considerable amount of
time, we have relied on credit scores to determine eligibility

IJISRT24FEB1225 www.ijisrt.com 1058

 Volume 9, Issue 2, February – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
for credit cards, but it's time for a change. This approach detection can be tackled using different methods, with
creates a distinction between those with and without, which is machine learning being just one of them. Sophisticated anti-
detrimental to business. Instead, the banking sector can gather fraud systems using financial machine learning can detect
data about each person's loan repayment behavior, the number subtle user behavior patterns and relationships. To detect
of loans they are currently managing, the number of credit potentially fraudulent transactions, the system can analyze
cards they have, and similar details. This data can help large amounts of data and evaluate different factors
customize the interest rate on a card to better align with the simultaneously and in real-time.
goals of the institution providing it. AI and ML play a crucial
role in this scenario, as they can analyze a large number of Using machine learning in payment processes provides
individual financial documents to generate a solution. AI, benefits to the payment industry. Owing to technology,
which relies on data and is driven by it, can analyze large payment service companies can reduce transaction costs,
amounts of data and make recommendations for loan credit leading to a rise in customer engagement. One of the
offers (Bauguess et al., 2017). advantages of using machine learning in payments is its ability
to enhance payment routing based on pricing, performance,
III. WHY USE MACHINE LEARNING AND AI IN functionality, and various other factors.
BANKING?
Machine learning systems can effectively allocate traffic
Machine learning and artificial intelligence enable to the best-performing variables by analyzing different sources
machines to perform a wide range of intricate tasks in place of of data. This ability allows financial institutions to deliver
humans. In a world where technology is prevalent in nearly optimal results to merchants according to their specific
every industry, financial institutions need to incorporate objectives. Today, there is a wide range of machine learning
advanced technology to stay competitive, improve their IT applications designed for use in the finance sector. These tools
systems, and meet current market needs. To elaborate on this, offer great potential for businesses to address common
here is the reason why ML and AI are essential for banking. challenges and generate significant value. Payment service
providers can use machine learning in payment processing to
The use of machine learning in banking is driving decide if a transaction should continue or if it needs to go
progress in the financial services industry (Analytics Vidhya, through a two-step verification process.
2019). With the help of innovative solutions, financial
institutions can now transform the constant flow of data they  Robo-Advisors (Portfolio Management)
generate into valuable insights for a wide range of Automated financial guidance and assistance is provided
stakeholders, including top executives, operational teams, by online tools known as robo-advisors. They provide
marketing, and business development. Companies are automated portfolio management services, using algorithms
increasingly using machine learning applications in finance to and data to create and manage a client's investment portfolio.
enhance security, user experience, and customer support, and These online investment platforms streamline the investment
streamline processing with minimal gaps. The combined process, easing the potential intimidation factor for some
advantages are so significant that AI and analytics could individuals. Furthermore, the cost of using these services is
potentially add up to a value of $1 trillion annually for the much lower compared to the fees of a financial advisor. In
global banking industry. addition, a significant number of them have very low or no
account minimum requirements. Betterment and Wealthfront
According to Analytics Vidhya (2019), machine learning are two digital investment companies that provide automated
involves analyzing large volumes of data to gain insights and portfolio management and financial advice through an app or
proficiency in performing specific tasks, such as online platform.
distinguishing counterfeit legal documents from authentic
ones. The finance industry generates a large amount of These companies are digital financial advisors that use
intricate and vast data, which machine learning is highly technology to help clients manage their money. Betterment
skilled at handling. The following are five ways in which the uses computer algorithms to suggest a fitting distribution of
banking and finance industry has been influenced by machine assets for investors. The conclusion is based on the investors'
learning. reactions to questions about their intended use of the funds
and their projected time frame. Wealthfront leverages the
 Anomaly Detection advantages of technology to provide its investment services in
Identifying anomalies in the asset-serving division of an impersonal manner. The software is programmed to
companies is extremely challenging. Mistakes or issues in implement established investment strategies, identify high-
everyday processes can lead to unusual occurrences. It is quality investment opportunities, and consistently manage the
important to detect any unusual or abnormal activities in the optimal investment portfolio. Nutmeg is one of the largest
fintech industry as they may be linked to illegal activities such online investment management companies in the UK. The
as account takeover, fraud, hacking, or money laundering, Nutmeg robo-advisor uses financial goals and risk tolerance to
ultimately leading to unforeseen consequences. Anomaly allocate funds into a diverse portfolio.

IJISRT24FEB1225 www.ijisrt.com 1059

 Volume 9, Issue 2, February – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
 Algorithmic Trading Traditional banks that are still relying on paper forms for
Algorithmic trading allows the execution of large new client onboarding can take advantage of machine learning
transactions by breaking them into smaller "child orders" and technology for processing documents. Machine learning is a
delivering them to the market at intervals. Hedge fund powerful and scalable method for onboarding, able to handle
managers are the primary users of machine learning in various types of data such as ID scans or invoices. Clients can
finance, as they also utilize automated trading systems. To quickly open a bank account and complete necessary
stay ahead of the competition, it allows traders to automate verifications in real time. Machine learning applications help
certain tasks. Additionally, the technology allows for cross- companies develop positive and enduring connections with
marketplace operation, improving opportunities for trading. their clients.
Another benefit for financial institutions that employ machine
learning is the ability of the algorithms to continuously adapt  Fraud Detection and Compliance
and respond to current market events, giving them a Fraud poses a significant challenge for financial
competitive edge. institutions, and it is one of the primary reasons for
implementing machine learning in the banking industry.
 Banking Processes That Use AI Machine learning systems can identify fraudulent activity by
Banks use machine learning technology for a range of utilizing different algorithms to analyze large amounts of data.
functions. Below are the most common benefits that banking Banks can oversee transactions, observe client behavior, and
and financial institutions receive from the use of ML and AI. record data for additional compliance and regulatory systems
to reduce overall risk in regulatory compliance. Fraud is
 Credit Scoring becoming a growing concern for businesses in the fintech
One of the most promising uses of machine learning in industry, regardless of their customer base or scale. Financial
banking is likely to be credit scoring (Ram, T, 2023). It machine learning can analyze large sets of simultaneous
evaluates a customer's financial capability and their likelihood transactions in real time. Simultaneously, ML has the potential
of devising a repayment strategy for their debts. There is a to decrease human involvement by continuously improving
critical demand for credit scoring solutions due to the large models and gaining knowledge from results.
number of unbanked individuals worldwide, with only about
half of the population meeting the requirements for credit. Financial institutions can use machine learning to
Various types of information such as work history, total identify and verify user behavior, as well as promptly address
earnings, financial transactions, and credit records are all cyber threats. Automated artificial intelligence is now widely
considered in the evaluation process for machine learning connected to global fraud detection (Ray, S., 2023) With the
scoring. It is a statistical and accounting-based mathematical ability to easily recognize patterns as irregularities, businesses
model. Hence, machine learning algorithms can produce credit are now capable of making intelligent predictions about fraud.
score evaluations that are more precise, responsive, and AI and machine learning have the potential to enhance the
customized, allowing a greater number of individuals to obtain precision of real-time approvals, resulting in an overall
credit. enhancement of regulatory compliance. Furthermore, by
saving financial institutions a substantial amount of money,
financial organizations can improve the accuracy and
Machine learning systems can assess borrowers in an efficiency of their operations.
unbiased manner, unlike human evaluators. Additionally,
machine learning in banking can help organizations eradicate In 2015, banks and retailers worldwide experienced a
biases based on gender, race, or other conscious or total of £16.74 billion ($2184 billion) in fraud losses on debit,
unconscious factors, allowing them to serve a broader range of credit, and prepaid payment cards, as reported by Bloomberg.
customers. ML in credit scoring has numerous benefits, such Valuing an investment requires a complex series of
as allowing customers to easily apply for loans from their calculations. The method involves collaborating with multiple
homes with just a couple of clicks. teams responsible for different aspects of investment asset
management, as well as product specialists and portfolio
IV. ONBOARDING AND DOCUMENT PROCESSING managers. These teams should consider different approaches
to investing. The solution to this problem lies in using
In the past, handling documents has required a machine learning to process large amounts of data from
significant amount of time and effort. Ultimately, machine various sources in real-time while also capturing and
learning can expedite the classification, labeling, and understanding biases and preferences related to risk tolerance,
processing of documents. Before machine learning algorithms investments, and time horizon.
can analyze the context of scanned documents, optical
character recognition (OCR) needs to be used on copies to  Personalized Offers
interpret the text. The machine learning model categorizes and The banking industry faces a major obstacle in the form
organizes data for future reference. of fraud, which is why the implementation of machine
learning is necessary. Machine learning systems use various

IJISRT24FEB1225 www.ijisrt.com 1060

Volume 9, Issue 2, February – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
algorithms for analyzing vast amounts of data and detecting deep learning, and big data analytics can have a beneficial
fraudulent activity. Banks can monitor transactions, observe effect, especially in the risk management sectors of credit,
customer activities, and maintain records for additional liquidity, market, operational risk, and related areas.
compliance and regulatory systems to minimize overall risk in
regulatory compliance. Fraud is an increasing worry for V. CONCLUSION
businesses in the fintech sector, regardless of their size or the
type of customers they serve. Financial machine learning can In conclusion, the fusion of Machine Learning (ML) and
process vast amounts of concurrent transactions instantly. At Artificial Intelligence (AI) marks a major change in the
the same time, ML can reduce human intervention by banking sector, signaling a fresh era of effectiveness,
constantly enhancing models and learning from outcomes. creativity, and customer focus. The far-reaching impact of
computational intelligence is clear in all areas of banking, as it
Financial institutions can utilize machine learning to improves decision-making and interaction with customers.
detect and confirm user activity and to quickly combat Although there were some initial concerns, financial
cybersecurity risks. Automated artificial intelligence has institutions are now realizing the essential role of AI and ML
become extensively integrated into global fraud detection, in staying competitive and relevant in the digital era.
allowing businesses to effectively identify irregular patterns
and make informed forecasts about potential fraudulent The combined efforts of traditional banks and fintech
activities (Ray, S., 2023). Artificial intelligence and machine companies are speeding up the process of transformation,
learning can improve the accuracy of immediate approvals, promoting an environment of creativity and adaptability in the
leading to an overall improvement in meeting regulatory industry. As we explore through the challenges of a changing
standards. Additionally, through cost savings for financial financial environment, it is crucial to continue to explore and
institutions, financial organizations can enhance the precision apply AI and ML technologies to maintain growth,
and effectiveness of their processes. adaptability, and customer happiness. By making strategic
investments and forming partnerships, banks can fully utilize
In 2015, Bloomberg reported that banks and retailers computational intelligence to bring about significant changes
globally suffered a combined loss of £16.74 billion ($2184 and influence the future of the banking industry.
billion) due to fraudulent activity on debit, credit, and prepaid
payment cards. Evaluating an investment involves a REFERENCES
sophisticated set of mathematical calculations. The approach
requires working closely with several teams overseeing [1]. Analytics Vidhya. (2019, June 25). 5 AI applications in
various aspects of investment asset management, along with Banking to look out for in the next 5 years.
product experts and portfolio managers. These teams ought to https://www.analyticsvidhya.com/blog/2017/04/5-ai-
explore alternative strategies for investment. The answer to applications-in-banking-to-look-out-for-in-next-5-years/
this issue is to use machine learning to analyze vast quantities [2]. Analytics Vidhya. (2019) What is Machine Learning?
of real-time data from different sources, while also identifying Understanding Machine Learning and its Types.
and understanding biases and preferences concerning risk https://www.analyticsvidhya.com/machine-learning/
tolerance, investments, and time frame. [3]. Arthur Samuel (1959). Some Studies in Machine
Learning Using the Game of Checkers. IBM Journal 3,
The financial industry, as well as the overall economy, (3): 210-229. https://ssrn.com/abstract=3226514
are being significantly impacted by the rising influence of [4]. Bauguess, Scott W., The Role of Big Data, Machine
artificial intelligence and machine learning. The significance Learning, and A.I. in Assessing Risks: A Regulatory
of artificial intelligence and machine learning in banking has Perspective (June 21, 2017). SEC Keynote Address:
piqued interest, especially in risk management, following the OpRisk North America 2017.
worldwide financial crisis. The research is centered on http://dx.doi.org/10.2139/ssrn.3226514
exploring how artificial intelligence and machine learning can [5]. Castelli, M., Manzoni, L., & Popovic, A. (2016). An
be used to enhance banking risk management. The paper aims artificial intelligence system to predict the quality of
to investigate the chances of successful implementation while service in banking organizations. Computational
also considering potential obstacles and their corresponding Intelligence and Neuroscience: CIN, 2016.
solutions. Artificial intelligence and machine learning can help http://dx.doi.org/10.1155/2016/9139380
address current global economic and financial difficulties, [6]. Das, S., Dey, A., Pal, A., & Roy, N. (2015). Applications
including those stemming from the COVID-19 pandemic. of artificial intelligence in machine learning: Review and
prospect. International Journal of Computer
This article primarily explores the management of credit Applications, 115(9) http://dx.doi.org/10.5120/20182-
risk, with a secondary focus on examining the use of artificial 2402
intelligence and machine learning in various other risk
management areas. It is determined that a strategic and well-
planned expansion of artificial intelligence, machine learning,

IJISRT24FEB1225 www.ijisrt.com 1061

Volume 9, Issue 2, February – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165
[7]. Donepudi, P. (2016). Influence of Cloud Computing in
Business: Are They Robust? Asian Journal of Applied
Science and Engineering, 5(3), 193-196.
https://doi.org/10.5281/zenodo.4110309
[8]. Kumar, A. S. & Chandrakala, D. (2016). A survey on
customer churn prediction using machine learning
techniques. International Journal of Computer
Applications, 154(10)
http://dx.doi.org/10.5120/ijca2016912237
[9]. Purdy M. & Daugherty, P. (2016). Why artificial
intelligence is the Future of Growth. Accenture,
https://www.accenture.com/t20170524t055435__w__/ca-
en/_acnmedia/pdf-52/accenture-why-ai-is-the-future-of-
growth.pdf
[10]. Ram, T. (2023, February 17). AI in Daily Life:
Applications and Threats. Analytics Vidhya.
https://www.analyticsvidhya.com/blog/2023/02/ai-in-
daily-life-applications-and-threats/
[11]. Ray, S. (2023, February 17). Top 10 AI & Data Science
Trends to Watch in 2023. Analytics Vidhya.
https://www.analyticsvidhya.com/blog/2023/02/top-10-
data-science-and-ai-trends-to-watch-in-2023/
[12]. SEC Speech, Has Big Data Made Us Lazy?, Midwest
Region Meeting of the American Accounting
Association, October 2016.
https://www.sec.gov/news/speech/bauguess-american-
accounting-association-102116.html
[13]. Taher-Uz-Zaman, M., Ahmed, M. S., Hossain, S.,
Hossain, S., & Jamal, G. R. A. (2014). Multipurpose
Tactical Robot. Engineering International, 2(1), 21-27.
https://doi.org/10.18034/ei.v2i1.204
[14]. Yu, L., Yang, Z., & Tang, L. (2016). A novel multistage
deep belief network based extreme learning machine
ensemble learning paradigm for credit risk assessment.
Flexible Services and Manufacturing Journal, 28(4), 576-
592. https://link.springer.com/article/10.1007/s10696-
015-9226-2

IJISRT24FEB1225 www.ijisrt.com 1062

View publication stats