Blog

Banking, Lending and Fintech Topics

on Sep 12th, 2018.

Fraud Fighting and AI


 

Security Pacific National Bank was the first FI to have used AI in 1087, although compared to today’s, it was pretty basic technology. In 2016 Mastercard introduced Decision Intelligence, the first AI network implemented on a global scale. We can also see IBM’s Watson being used in several industries, from health to finance. Even big banks, such as HSBC utilize AI in their everyday operations, such as AML/KYC.

 

How does AI work?

 

This rapid adoption of AI and machine learning technology is due to the ability to process enormous amount of data in a short time. An AI machine can sort through raw data and give results in just a few hours. They use predictive rules that recognize anomalies, thus minimizing “false positives” which increases customer experience. 

 

In 2015 a research conducted by Javelin Strategy found that these false positives costed retailers around $118 billions annually. Lost customers are not calculated in this number, which is estimated at around of one third of non-returning customers and around 40% stopped using their card after being falsely declined. 

 

The old way of fraud assessment was, simply said, a checklist with pre-defined binary rules that had to be manually reviewed. 

 

HSBC brought in AI into AML after an $1.9 billion settlement followed charges for allegedly laundering $881 million for Mexican drug cartels. The AI software HSBC integrated from Quantexa is based on scanning Big Data to understand the flow of money. The software scans for s phone numbers, addresses, company directors and news reports. The other software from Ayasdi, which HSBC integrated to automate parts of their AML investigations is based on clustering - a technique that segments customers into fine-tuned clusters based on their behaviour. Each customer behaviour is compared against the clusters’ own level of risk to lower the number of false negatives. If customers don’t fall into any of these clusters, they would be considered suspicious right away.

 

Mastercard’s Decision Intelligence, for example, “examines how a specific account is used over time to detect normal and abnormal shopping spending behaviors. In doing so, it leverages account information like customer value segmentation, risk profiling, location, merchant, device data, time of day, and type of purchase made.” 

 

Since the launch of Decision Intelligence in 2016, Mastercard grew revenu by 23%. 

 

Many FIs invest in AI in order to improve their customer service and performance and increase revenue.

 

Will AI replace humans?

 

As things are and are certainly taking a direction, many fear that AI will replace the humans and leave people jobless. 

 

Truth be told, AI will replace the manual work just like any new technology did in the past. But this doesn’t mean that people will be left jobless. Not so long ago, electricity and the light bulb profoundly changed the lives of people. 

 

 “I truly believe AI is going to replace many of the activities humans do,” said Omar Almaraz, Head of KYC at Dolare. “Nevertheless, humans activities tend to evolve as well,” he continues, “when cars appeared they replaced many jobs, however they also created a whole industry.”

 

Maybe AI did decrease the number of people required to review data and investigate transactions, but it did create a whole new industry. Many fintech companies deal with AI and machine learning tech, and these companies don’t just need engineers and developers, they also need investigators’ experience and expertise. Now, investigators are migrating from executive positions to consultant, and the employment demand is increasing proportionally with the number of companies developing AI to fight fraud. 

 

Although, AI has great potential, the adoption in the fraud fighting is slow. This is in part as a result of insufficient understanding of how AI operates. Compliance officers are required to certify the outputs and the processes of the AML models, meaning that the lack of understanding how AML AI software operates is an obstacle. However, as any rising technology there is a period of adaptation, and we can see that  AI is being developed and applied in the field of fraud fighting. The AI holds great potential which can benefit all, stakeholders, customers, and regulators.

 

Sources:

  1. http://finovate.com/feedzai-brings-machine-learning-to-the-fraud-fight/
  2. https://www.intel.com/content/www/us/en/financial-services-it/article/insurance-fraud-revealed.html
  3. https://www.businessinsider.com/could-ai-be-the-ultimate-weapon-in-the-fight-against-fraud-2017-1
  4. https://hbr.org/2018/08/the-risks-and-benefits-of-using-ai-to-detect-crime?utm_campaign=ACFCS%20Daily%20Brief&utm_source=hs_email&utm_medium=email&utm_content=65203647&_hsenc=p2ANqtz-_uvOwaopRubnVgMAhOc_IQ1M0O1c_5CCB3WJc9Mggv3Nxh1GNRlt8ijVBTm_LtP7ArJIpFFIghA2cwlqzkIdHznw_JAg&_hsmi=65203647
  5. http://www.acfe.com/fraud-examiner.aspx?id=4294999437
  6. https://www.forbes.com/sites/theyec/2018/06/04/artificial-intelligence-and-the-future-of-financial-fraud-detection/#4e9ceb95127a
  7. https://sigmoidal.io/real-applications-of-ai-in-finance/
  8. https://www.businessinsider.com/mastercard-artificial-intelligence-fraud-protection-2017-1?IR=T
  9. https://newsroom.mastercard.com/press-releases/mastercard-rolls-out-artificial-intelligence-across-its-global-network/
  10. https://www.techemergence.com/machine-learning-in-finance/
  11. https://www.finextra.com/blogposting/14748/ai-implementation-in-aml-at-hsbc-sees-a-considerable-reduction-in-compliance-costs
  12. https://www.ft.com/content/b9d7daa6-3983-11e8-8b98-2f31af407cc8
  13. https://www.forbes.com/sites/steveculp/2018/01/16/are-artificial-intelligence-and-machine-learning-the-next-frontiers-for-fighting-money-laundering/#e88c41b4a639
  14. http://www.bobsguide.com/guide/news/2017/Sep/21/how-hsbc-is-using-ai-in-its-anti-money-laundering-compliance/