AI and the hope of what it can deliver
Peter Marini, Sales Director, Asia Pacific at NetGuardians talks with Hubbis CEO Michael Stanhope about where artificial intelligence, or AI, stands today in its ability to meet exigencies within the financial industry when it comes to the detection and prevention of financial fraud.
The cyber heist of Bangladesh Bank in 2016, in which malware was used to fraudulently withdraw USD 1 billion from the bank, was only one of a string of attacks over the last decade or so by hackers who spend multiple years planning attacks on banks, in almost all instances using the SWIFT global payment network.
Banks and financial institutions are increasingly looking to AI and expecting that the technology will be able to deliver the ‘holy grail’ of protection when it comes to fraud detection, investigation and prevention, whether in current cases of fraud or simply ‘to stay ahead of the curve’ when it comes to cyber criminal activity. But does AI technology truly minimise or eliminate those requirements for the bank?
Marini explains that AI has two different models – supervised and unsupervised learning – and says that users need to leverage both types of technology to make the most of AI, along with other methods of fraud prevention.
To do that, Marini says, there is a need to understand when an instance of fraud has occurred within the organisation, and point the software to the occurrence.
“You will always need human interaction to dictate what is truly fraud or just suspicious behaviour, but for an appropriate reason,” Marini says. ”Maybe the transaction is just someone withdrawing money on regular basis to accumulate enough cash to buy something they can’t afford to make one payment on,” he adds. “These are the things a human investigator would be able to notice, where a machine possibly would not.”
He cautions that the current reality is that while AI can deliver a lot of the safeguards when it comes to fraud detection and prevention, the technology will never eliminate the need for human intervention and investigation.
For Asian wealth management firms and their strategies around fraud prevention, Marini says a combination both AI and human investigators equipped with ‘augmented intelligence’ is required to reduce the number of false positive detections.
“We need AI to leverage the algorithms to reduce the number of false positive detections, and we need human investigators to review the appropriate fraudulent transactions, which are those that are unusual or different, and therefore difficult for machines to detect.”
More from Peter Marini, NetGuardians