Earlier this year, our colleague Roy Prayikulam had the opportunity to present at the Merchant Payments Ecosystem (MPE) conference in Berlin, where INFORM participated as a silver sponsor. The goal of Roy’s presentation was to provide the audience with a better understanding of how artificial intelligence (AI) and machine learning can help prevent fraud and reduce the risk of payment defaults.
At the beginning, after demonstrating AI has been around for decades, Roy asks the question, “Why are we experiencing such a big hype around AI now?” The answer is that we now have the right ingredients to cook up something special; those ingredients include accessibility to affordable computing power and the availability of data.
What does this mean for the payment ecosystem? Due to the wide usage of AI in everyday life, such as spam filters, word prediction in text messages and personal assistants such as Siri, consumer expectations are changing. We are expecting our service providers to enable us to use these new technologies. As an industry specific example, Roy mentions payment authentication while making a purchase. With all the available tech today, we will likely not be carrying around long passwords or tokens in the coming years, but instead using biometrics and other innovative forms of authentication.
The 3 main takeaways from the presentation:
Yes, machine learning is valuable for fraud detection in the payment ecosystem. However, one size doesn’t fit all: Pick your algorithms wisely. And lastly, data isn’t everything – don’t forget the human element and the knowledge we have gathered over the years.
Find out more about INFORM’s Hybrid AI approach to payment fraud prevention and risk management in the video below:
Brief introduction of AI (1:41)
What do AI and Machine Learning mean for the payment ecosystem (4:50)
Using machine learning to address new challenges (7:15)
Machine learning in the context of AI (8:10)
Supervised vs. unsupervised learning (9:17)
Supervised learning in the context of fraud prevention and risk management (12:33)
Unsupervised learning in the context of fraud prevention and risk management (15:30)
Machine learning is data driven (16:43)
Hybrid AL explained (18:17)
3 tips to consider when using machine learning against fraud (20:45)
Many thanks to Filip Rasovsky and the MPE team for making the video of this presentation available to us.