Detecting and Preventing Banking Application Fraud: Uncovering the

True Challenge of Synthetic Identities with Analytics and Artificial

Intelligence

Is the financial services industry facing a huge issue? In a recent case

prosecuted by the US Department of Justice, the financial services

industry faced a staggering challenge: approximately 7,000 fake identities

and 25,000 fraudulent credit cards obtained through banking application

fraud. The scheming fraud caused a massive problem in the business

world that affected businesses and resulted in more than $200 million in

losses.

Synthetic identity fraud is a form of deception in which criminals

manipulate a combination of real and fabricated credentials to create

entirely new identities devoid of any genuine connections to actual

individuals. These counterfeit personas are meticulously crafted by

inventing entirely fictional information, piecing together details from

various sources, or exploiting data compromised through security

breaches.

To combat this growing threat, the financial services industry has turned to

advanced analytics and artificial intelligence. These technologies play a

pivotal role in identifying and preempting instances of synthetic identity

fraud. In this review, we will examine the intricacies of this issue, evaluate

the proposed solutions, and provide critical insights into the current state

of the financial services sector.

Moving forward, it is crucial to delve deeper into the implications of the

article's findings. In the article ‘Detecting and Preventing Banking

Application Fraud: Uncovering the True Challenge of Synthetic Identities

with Analytics and Artificial Intelligence, early detection can help locate

and prevent losses in the financial services industry. As stated, with the

collaboration of SAS and Asia Pacific Bank, they aid in locating scheming

fraud in the form of 60,000 contact phone numbers referencing

immigration agents and 5,000 contact numbers referencing casinos. 2,500

phone numbers referencing the bank branch at which the application was

made; 1,500 numbers referencing a meat processing plant. Through these

activities, the fraudster can easily do their work; however, using SAS

software, the bank found four times more banking application fraud,

valued at $3 million a month, compared to its former techniques. Which

means that with the help of SAS software that provides the most complete

and sophisticated data management capabilities, fraud can be avoided,

resulting in no loss of money.

Furthermore, the article highlights that the issue often begins with the theft

of identities or the manipulation of credentials but evolves into synthetic

identity fraud. Synthetic identity fraud involves the creation of fabricated

credentials that are not linked to real individuals, making it a challenging

form of fraud to detect, Gartner estimates that synthetic identities are

responsible for 20 percent of credit charge-offs and a staggering 80

percent of credit fraud losses. This fraudulent practice involves several

steps, including applying for credit, adding the synthetic identity as an

authorized user on existing accounts, and collaborating with merchants to

fabricate credit accounts and reports. The pervasive nature of banking

application fraud becomes evident through this intricate web of deceit.

Moreover, the article delves into the stages of banking application fraud

and discusses how analytics and machine learning can be effectively

applied at each stage to detect fraudulent activity. These stages include

monitoring application data, assessing past experiences, analyzing social

networks, and tracking transaction patterns. For instance, monitoring

application data to identify reused information or devices across multiple

identities that appear unrelated is crucial. Additionally, searching for

'proof of life,' which includes comprehensive details about the identity,

such as driver's licenses, voter registrations, or property ownership, can

help in fraud detection. Monitoring payments from the same source (bank
and account) for otherwise unrelated accounts and observing increased

transaction frequency are also valuable indicators.

To provide some basis for why artificial intelligence is better in fighting

against synthetic fraud than humans, I have cited some articles to prove

that artificial intelligence is better than the traditional process. The

traditional process includes the manpower to locate the problem or issue in

the financial services industry that possibly cannot detect any fraud. Also,

I provide an example of an artificial intelligence algorithm that can fight

fraud.

According to Eray Eliaçık (2022), when it comes to analyzing data,

computers outperform humans by a large margin. That is why artificial

intelligence (AI) and machine learning (ML) offer immense potential.

They provide us with the ability to examine and make sense of intricate

patterns that are beyond the scope of human comprehension. These

technologies empower us to identify and act upon insights that would

otherwise remain elusive to the human mind. Let us discuss a few of the

methods in this paper that are used for preventing fraud.

AI algorithms take out crucial details from the data gathered in the historic

data lake. These details, called features, can be things like the amount of a

transaction, the time it happened, how users behave, where it took place,
and information about the device used. Feature engineering is the process

of picking and changing these features to make them more useful for

spotting fraud. The AI model is taught using special data that has been

marked to show if it is a real or fake activity. This data includes examples

of things that are normal and things that are fraudulent. The model uses

machine learning algorithms, like supervised learning or anomaly

detection, to learn from this data. While it learns, the AI model figures out

the patterns and connections between the various parts of the activity data.

This helps it understand how to tell apart the fraudulent activities from the

regular ones.

After the AI model has been learned, it can be used to give fraud scores

for new transactions or activities. The model looks at the various parts of

the new data and gives it a number that shows how likely it is to be

fraudulent. This number is the fraud score. The fraud score can be a

number that keeps changing, or it can be a simple answer like "yes" or

"no" to show if it is a high or minimal risk for fraud.

According to Gary S. Reynolds, a massive portion of online activities can

be conducted through mobile devices using applications, and biometric

and facial recognition technologies have emerged as crucial authentication

methods. These technologies rely on sophisticated algorithms that

compare the stored biometric templates, such as fingerprint, facial, or iris

templates, with the input biometric data to determine a similarity score. If

the calculated score surpasses a predefined threshold, the biometric data is

deemed a match. Template matching algorithms employ diverse

techniques, including correlation based matching or Hamming distance, to

facilitate this comparison process.

To my account as a computer science student, this article talks about the

idea of fighting back against the criminal, especially in synthetic identity

fraud, which has become a global issue that affects the business world. It

becomes one of the reasons why some businesses and individuals who

become victims of this case lose their wealth, which they worked for their

whole lives. With the help of the advancement of analysis and artificial

intelligence that detects possible fraud and incoming fraud, which prevents

any transaction from being done and results in major losses, I firmly stand

that the advancement of analytics and artificial intelligence never become

a hindrance in business, hence it gives the reassuring help to individuals

and businesses to manage their accounts, detect fraud, and become a

security. In addition, this article emphasizes the importance of early

detection and advanced analytics in preventing banking application fraud.

It provides valuable insights into the methods used by fraudsters and the

strategies employed by financial institutions to combat this pervasive

issue. Analytics and Artificial Intelligence (AI) play a crucial role in

stopping fraud. These high-tech tools give organizations strong weapons

to fight against fraud, which keeps changing all the time. With the help of

AI algorithms, organizations can look closely at a lot of data and find

tricky patterns and strange things that might mean fraud. Artificial

Intelligence keep learning, so organizations can stay ahead of fraudsters

who keep coming up with new tricks. Fighting fraud is a never-ending job,

and new tech will keep showing up to help. Staying watchful, flexible, and

creative is key to staying ahead of the sneaky tricks used by fraudsters.

REFERENCES:

Barta, D. (2018, November 18). Detect and prevent banking application

fraud: Analytics and artificial intelligence uncover the real challenge –

synthetic identities. SAS Global Security Intelligence Practice. Retrieved

September 10, 2023, from

https://www.sas.com/en_ph/insights/articles/risk-fraud/detect-prevent-

banking-application-fraud.html

Eray Eliaçık, Aritificial Intelligence vs. Human Intelligence: Can a Game-

changing Technology Play the Game, 2022. [Online] Available:

https://dataconomy.com/2022/04/20/is-artificial-intelligence-better-than-

human-intelligence/

Imane Sadgali, Nawal Sael, and Faouzia Benabbou, “Human Behavior

Scoring in Credit Card Fraud Detection,” IAES International Journal of

Artificial Intelligence, vol. 10, no. 3, pp. 698-706, 2021

Data Robot. [Online]. Available: https://www.datarobot.com/wiki/feature/

Gary S. Reynolds, “Facial Recognition: A Biometric for The Fight

Against Check Fraud,” Journal of Economic Crime Management, vol. 4,

no. 2, 2006.

Gupta, P. (2023). Leveraging Machine Learning and Artificial Intelligence

for Fraud Prevention. Principal Data Engineer, Discover Financial

Services, USA. Retrieved from

https://www.researchgate.net/profile/Pankaj-Gupta-76/publication/371445

997_Leveraging_Machine_Learning_and_Artificial_Intelligence_for_Frau

d_Prevention/links/6483a75179a72237651b2e6c/Leveraging-Machine-

Learning-and-Artificial-Intelligence-for-Fraud-Prevention.pdf