NetGuardians’ Roy Belchamber was one of the industry experts on a panel at the recent AI Fraud and AML Summit in London. The panel discussed regulation, data-sharing to get the best out of AI models and how it was time to move from fraud detection to fraud prevention
The whole panel agreed: artificial intelligence (AI) is a game-changer when it comes to spotting and stopping fraud, bringing greater efficiency and speed. But there are significant challenges, including regulation, data-sharing and building models that are accurate without impeding the customer experience.
Andrew Churchill, cyber-crime expert and panel moderator, kicked off the discussion commenting that legislation lags behind technology when it comes to AI, particularly on data-sharing cross-border and privacy. In response, NetGuardians’ Roy Belchamber pointed out that the industry needs harmony when it comes to regulations and AI, but added that regulators were increasingly looking at this area.
“I think this is set to pick up because as a tool against fraud, AI is very effective. It can’t be ignored, and regulators are starting to tackle the issues,” he said.
Given the restrictions on cross-border data-sharing, some of the answer, certainly in the short term, is to share data locally.
Roy told the conference that NetGuardians had started a Community Scoring and Intelligence service that delivers significant improvements in the race to prevent financial crime. “A consortium of banks and partners that use NetGuardians software has started to share intelligence through us. This has led to a significant uplift in the performance of our models and enabled our customers to make better decisions when investigating suspicious cases,” he said.
The discussion also touched on the need for transparent or explainable AI models to ensure optimum performance and prevent biases from building up. It then moved on to the need to switch the current focus from fraud detection to fraud prevention without adding friction.
The key to this is building accurate profiles of customers so anything unusual stands out, raising an alert that can be used to suspend a payment before any money has left the bank.
“This [profile-building] starts with the client relationship. The bank builds up the profile and uses it to catch the fraud before it’s complete. In real time. You get one bite of the cherry to do this. The tools are there. We just incorporate data from upstream, various metrics and know-your-customer information, and we can accurately spot fraud attempts. We have effective models that don’t murder the customer experience and banks using our software are already putting prevention first.”
The session ended by revisiting regulation. This time, however, it touched on the speed at which banks need to comply with new rules and that all banks involved in a transaction need to have clear sight of risk along the whole transaction journey.
As an example of this, Roy brought up new regulations that require banks to start monitoring incoming payments for money laundering from December this year. “This is indicative of how fast banks need to react. The emphasis here is moving to the receiving bank and they need to comply by the end of the year. These are new, faster timescales,” he said.
The panel agreed that while this is tough, tackling financial crime and money laundering from both sides would bring better protection to banks from the criminals when it comes to fraud and money laundering.
Roy Belchamber is head of product management at NetGuardians
Find out more about NetGuardians’ AI-based solutions for fraud and financial crime prevention here.
Curious to see
Get a free demo: