Getty
Credit cards may be a convenient way to spend hard-earned money, but they can also make for a convenient way for thieves to steal said money. Credit card numbers can be stolen without your knowledge. Until you spot a fraudulent charge on your monthly statement, you may have no idea your information has been stolen. Credit cards can be stolen in a variety of ways: Through theft of a physical card, via data breaches, by card skimmers the list goes on…..Story continues…
By: Chauncey Crail & Dylan Pearl
Source: Forbes Advisor
.
Critics:
Credit card fraud is an inclusive term for fraud committed using a payment card, such as a credit card or debit card. The purpose may be to obtain goods or services or to make payment to another account, which is controlled by a criminal. The Payment Card Industry Data Security Standard (PCI DSS) is the data security standard created to help financial institutions process card payments securely and reduce card fraud.
Credit card fraud can be authorised, where the genuine customer themselves processes payment to another account which is controlled by a criminal, or unauthorised, where the account holder does not provide authorisation for the payment to proceed and the transaction is carried out by a third party. In 2018, unauthorised financial fraud losses across payment cards and remote banking totalled £844.8 million in the United Kingdom.
Whereas banks and card companies prevented £1.66 billion in unauthorised fraud in 2018. That is the equivalent to £2 in every £3 of attempted fraud being stopped. Credit card fraud can occur when unauthorized users gain access to an individual’s credit card information in order to make purchases, other transactions, or open new accounts. A few examples of credit card fraud include account takeover fraud, new account fraud, cloned cards, and cards-not-present schemes.
This unauthorized access occurs through phishing, skimming, and information sharing by a user, oftentimes unknowingly. However, this type of fraud can be detected through means of artificial intelligence and machine learning as well as prevented by issuers, institutions, and individual cardholders. According to a 2021 annual report, about 50% of all Americans have experienced a fraudulent charge on their credit or debit cards, and more than one in three credit or debit card holders have experienced fraud multiple times.
This amounts to 127 million people in the US that have been victims of credit card theft at least once. Regulators, card providers and banks take considerable time and effort to collaborate with investigators worldwide with the goal of ensuring fraudsters are not successful. Cardholders’ money is usually protected from scammers with regulations that make the card provider and bank accountable. The technology and security measures behind credit cards are continuously advancing, adding barriers for fraudsters attempting to steal money.
There are two kinds of card fraud: card-present fraud (not so common nowadays) and card-not-present fraud (more common). The compromise can occur in a number of ways and can usually occur without the knowledge of the cardholder. The internet has made database security lapses particularly costly, in some cases, millions of accounts have been compromised. Stolen cards can be reported quickly by cardholders, but a compromised account’s details may be held by a fraudster for months before any theft, making it difficult to identify the source of the compromise.
The cardholder may not discover fraudulent use until receiving a statement. Cardholders can mitigate this fraud risk by checking their account frequently to ensure there are not any suspicious or unknown transactions. When a credit card is lost or stolen, it may be used for illegal purchases until the holder notifies the issuing bank and the bank puts a block on the account. Most banks have free 24-hour telephone numbers to encourage prompt reporting. Still, it is possible for a thief to make unauthorized purchases on a card before the card is cancelled.
Card information is stored in a number of formats. Card numbers – formally the Primary Account Number (PAN) – are often embossed or imprinted on the card, and a magnetic stripe on the back contains the data in a machine-readable format. Fields can vary, but the most common include the Name of the cardholder; Card number; Expiration date; and Verification CVV code. In Europe and Canada, most cards are equipped with an EMV chip which requires a 4 to 6 digit PIN to be entered into the merchant’s terminal before payment will be authorized.
However, a PIN is not required for online transactions. In some European countries, buyers using a card without a chip may be asked for photo ID at the point of sale. In some countries, a credit card holder can make a contactless payment for goods or services by tapping their card against a RFID or NFC reader without the need for a PIN or signature if the cost falls under a pre-determined limit. However, a stolen credit or debit card could be used for a number of smaller transactions prior to the fraudulent activity being flagged.
Card issuers maintain several countermeasures, including software that can estimate the probability of fraud. For example, a large transaction occurring a great distance from the cardholder’s home might seem suspicious. The merchant may be instructed to call the card issuer for verification or to decline the transaction, or even to hold the card and refuse to return it to the customer. Given the immense difficulty of detecting credit card fraud, artificial and computational intelligence was developed in order to make machines attempt tasks in which humans are already doing well.
Computation intelligence is simply a subset of AI enabling intelligence in a changing environment. Due to advances in both artificial and computational intelligence, the most commonly used and suggested ways to detect credit card fraud are rule induction techniques, decision trees, neural networks, Support Vector Machines, logistic regression, and meta heuristics. There are many different approaches that may be used to detect credit card fraud.
For example, some “suggest a framework which can be applied real time where first an outlier analysis is made separately for each customer using self-organizing maps and then a predictive algorithm is utilized to classify the abnormal looking transactions.” Some problems that arise when detecting credit card fraud through computational intelligence is the idea of misclassifications such as false negatives/positives, as well as detecting fraud on a credit card having a larger available limit is much more prominent than detecting a fraud with a smaller available limit.
One algorithm that helps detect these sorts of issues is determined as the MBO Algorithm. This is a search technique that brings upon improvement by its “neighbor solutions”. Another algorithm that assists with these issues is the GASS algorithm. In GASS, it is a hybrid of genetic algorithms and a scatter search. Touching a little more on the difficulties of credit card fraud detection, even with more advances in learning and technology every day, companies refuse to share their algorithms and techniques to outsiders.
Additionally, fraud transactions are only about 0.01–0.05% of daily transactions, making it even more difficult to spot. Machine learning is similar to artificial intelligence where it is a sub field of AI where statistics is a subdivision of mathematics. With regards to machine learning, the goal is to find a model that yields that highest level without overfitting at the same time. Overfitting means that the computer system memorized the data and if a new transaction differs in the training set in any way, it will most likely be misclassified, leading to an irritated cardholder or a victim of fraud that was not detected.
The most popular programming used in machine learning are Python, R, and MatLab. At the same time, SAS is becoming an increasing competitor as well. Through these programs, the easiest method used in this industry is the Support Vector Machine. R has a package with the SVM function already programmed into it. When Support Vector Machines are employed, it is an efficient way to extract data. SVM is considered active research and successfully solves classification issues as well.
Playing a major role in machine learning, it has “excellent generalization performance in a wide range of learning problems, such as handwritten digit recognition, classification of web pages and face detection.” SVM is also a successful method because it lowers the possibility of overfitting and dimensionality.
Leave a Reply