Post2

AI in Banking – Fraud Detection

Posted by Samar Naqvi

Hello readers, as promised, we are here with our third blog in the series “AI in Banking and Financial Services”. While our previous two blogs focused on Anti-Money Laundering (AML) Pattern Detection and Chatbots respectively, in this blog, we are focusing on the impact of AI in Fraud Detection.

Frauds in Banking and Financial Services

Earlier, customers had limited channels available to interact and transact with their banks, which is why even the frauds in the banking industry were limited to loan defaults and robbery. The proliferation of technology, however, opened up numerous channels of interactions and transactions – website, mobile application, and ATM – which has significantly increased the instances and modes of fraud. From hacking to Ponzi schemes to phishing, daily there are reports of financial frauds in some part of the world.

Recent Global Financial Frauds:

  • In 2016, a bank in Bangladesh was robbed of $81 Million by some hackers by fraudulently manipulating bank’s SWIFT system.
  • In 2015, hackers stole about £650 million from about hundreds of global financial institutions by gaining access to secure information.
  • In 2009, the FBI cracked the largest phishing case ever when the court charged the US and Egyptian fraudsters for using phishing scams to steal account details of hundreds and thousands of people to transfer about $1.5 million into fake accounts they controlled.
  • Unearthed in 2008 the Bernie Madoff Scandal is believed to be the mother of Ponzi schemes. For over 20 years, Bernie Madoff successfully defrauded thousands of investors of billions of dollars ($64.8 billion as per Federal investigations) through his investment strategy called “split-strike conversion.”

The above examples are just the tip of the iceberg, there are many yet to be uncovered. The Federal Reserve believes global banking to be in line for $526 Billion in losses during the next crisis, if it occurs.

Role of AI in Fraud Detection

The advent of technology has made it easy for fraudsters to execute their plans. It can be argued that experienced bankers are armed with intelligence to detect and counter fraudulent transaction, which is not possible with machines.

This is where Artificial Intelligence (AI) and Machine Learning (ML) come into the picture. These next-gen technologies are able to assist humans in detecting patterns and taking judgment calls. The need for banking and financial service industry (BFSI) is to tap into the complete potential of AI including Machine Learning (ML), Natural Language Processing (NPL), Natural Language Understanding (NLU), Artificial Neural Networks (ANN), and Pattern Recognition to not only detect frauds but also catch the fraudsters.

Let us understand the role of AI in fraud detection by recognizing the types of frauds that have plagued the industry and how they are conducted.

Identity Theft

In the case of identity theft, an unauthorized person captures the id and password of a customer and uses it to log into the customer’s account. The need here is to ascertain that the person logging into the account is the bonafide customer.

Implementing AI to fight off identity theft involves integrating biometric identification systems, such as voice and facial recognition, into the login module to strengthen the identity verification process.

Fraudulent Transaction

Another concern for banks is to prevent fraudulent transactions, especially in the case of online funds transfer. Banks need a system that raises an alarm the moment it detects any irregular activity. With AI & ML supporting self-learning systems, banks can use these technologies to understand customers’ transaction patterns. If the transaction is in line with earlier transactions, it will not be labeled suspicious, however, if the transaction involves transfer of unusually high funds to a new party, it can be labeled suspicious and an alert can be raised.

 ‘Functioning of a typical system based on neural network technique’

Reference: http://blog.citizennet.com/blog/2012/11/10/random-forests-ensembles-and-performance-metrics

In the End

Going by the examples or cases highlighted, AI and ML can serve the BFSI in detecting frauds and catching fraudsters. Implantation of the technology, however, is an overwhelming and costly affair. It involves modernizing the traditional systems, training the personnel, and helping customers adapt to the change. Although the gains of implementing AI outweigh the cost, time, and adaptation needs. In addition to detecting frauds, integrating AI within banking functions enables banks to gain better insights into their customers’ expectations and preferences, and accordingly personalize their services.

Do let us know your thoughts about AI’s role in fraud detection in the comments below.

In our next installment, we shall be covering the next topic ‘AI in Stock Trading’.

Related Posts

  • AI in Banking(2) – ChatbotsAI in Banking(2) – Chatbots

    Hello readers, as promised, we are here with our 2nd blog in the series of blogs on “AI in Banking and Financial Services”, focusing on Chatbots. As discussed in the…

  • AI in Banking and Financial Services (AML)AI in Banking and Financial Services (AML)

    Recently HSBC announced that it is implementing AI technology to automate AML investigations - (Source: Reuters 01-June-2017) Similarly, many more banks have started using AI technology in various fields. Since…

  • All You Need to Know About BlockchainAll You Need to Know About Blockchain

    Blockchain is quite the buzzword these days, but how many of us really know about it? Don’t fret if you don’t, this blogpost will prove to be a great source…

  • Introduction to Chat bots (Part 1)Introduction to Chat bots (Part 1)

    Chat bots Chat bots are Conversational agents or a Dialog system that simulates Intelligent conversation using text or speech. Why chat bots? In recent times, a sudden rise in interest towards…

  • Mobile App Automation Testing using ‘ESPRESSO’

    If you are a Mobile Apps Test Engineer, you cannot overlook the very reliable Google Product i.e. Espresso. Espresso is an automatic UI testing or as we call it “hands…

  • Content Data Store

    Content Data Store Content Data Store (CDS) is a system to provide storage facilities to massive data sets in the form of images, pdfs, documents and scanned documents. This dataset…

Leave a Reply

Your email address will not be published. Required fields are marked *