Have you been victimized and tricked with an authorized push payment?

I’ve been reading recently about fraud, specifically banking fraud. Here at HPE this is also Cybersecurity awareness month – which means we all have to take training to avoid phishing and other security issues at HPE. The training is not too long and, I think, worth the effort to remind us to be vigilant. It certainly doesn’t take any time to click on a link by mistake.

HPE also sends out “test” phishing emails. If you accidently click on it – more training. If you forward it to our “report suspicious” link you are congratulated. It’s the price everyone pays nowadays to protect ourselves online. Recently I was reading about push payment fraud.

In an authorized push payment (APP) fraud, customers are tricked into transferring money. Fraudsters pose as a bank representative. They claim your account has been compromised, pressuring you to transfer funds into a different account.

These manipulations are made even more effective by advanced GenAI technologies like deepfakes, voice cloning, and SMS/phishing scams. The expansion of these technologies threatens to make common APP fraud tactics more realistic and puts consumers at much greater risk.

What can be done to prevent APP Fraud?

To fortify defenses against APP fraud, banks can take several proactive steps:

• • Customer Education: Provide customers with comprehensive education on APP fraud methods, emphasizing the importance of verifying payment requests and being wary of social engineering. Banks can’t require training such as the kind I mentioned within HPE but education can help.
  • Stronger Verification: Implement multi-factor authentication and enhanced due diligence for high-risk transactions or changes to payee details. Banks can’t force multi-factor authentication but it should be STRONGLY suggested even though it requires extra effort.
  • Real-Time Monitoring: Utilize real-time transaction monitoring systems to identify suspicious transactions and delay or block them when necessary.
  • Confirmation Protocols: Establish protocols to confirm high-value or unusual transactions with customers through separate channels, such as text-message verification of the transaction. I’m experiencing this more and more with my bank.
  • Know Your Customers’ Normal Patterns: AI and machine learning can detect deviations like increased transfers to new accounts or unusually high amounts by analyzing transfer frequency, recipient history, and transaction values.
  • Invest in GenAI-based Solutions: Banks shouldn’t cede ground to criminals on GenAI adoption. Invest in GenAI agents that can catch scams in progress and empower customers to recognize the warning signs for themselves. If criminals are using AI (and they are), banks need to be using it also.

I’m hoping many of you attended the TBC (Technology and Business Conference – rebranded (pun intended) from the Technology Boot Camp) in Houston. It was an excellent conference, themed around Houston and Texas. If you attended the party night, you would have received a customized cowboy hat. More on TBC in a later article.

I did want to highlight one aspect of Casey Taylor’s (HPE Nonstop VP & GM) keynote regarding Nonstop and AI. She mentioned the use of adjacent AI with Nonstop. I was on a taskforce a few years back to determine if there was any value in putting GPU’s in the Nonstop processors. Our team concluded that there wasn’t since our processing sweet spot is transaction processing and message switching. Nonstop processing compute intense workloads didn’t seem to be an area in which we could excel, and there was a lot of competition in that space already. However, adjacent AI, such as some of the world-class HPE AI systems aligned with fraud detection and perhaps existing on our fabric makes a lot of sense.

Real-time fraud detection could be achieved across our high-speed fabric letting the AI system perform the massive number crunching. As AI systems get better through learning, false positives should decrease. I believe a close association with AI (adjacent AI) is in Nonstop’s near future.