IBM is leveraging the capabilities of real-time AI to assist banks in analyzing transactions for fraud on a large scale.
Credit card fraud remains a significant issue globally, resulting in losses of $32 billion for the industry last year. Banks have faced challenges in addressing this problem due to the difficulty of running deep-learning models at scale in real-time, often only applying detection models to less than 10% of high-volume transactions.
IBM claims to have addressed this challenge with the introduction of IBM z16, which integrates AI inferencing through the IBM Telum Processor with robust and secure high-volume transaction processing. This technology can handle 300 billion inference requests per day with only one millisecond of latency.
For consumers, this advancement could mean a quicker and less cumbersome process for managing fraudulent transactions on their credit cards. For merchants and card issuers, it may lead to decreased revenue loss as customers face fewer false declines.
Additionally, IBM z16 is recognized as the industry’s first quantum-safe system, built on lattice-based cryptography.
Ric Lewis, Senior Vice President of IBM Systems, stated, “IBM is the gold standard for highly secured transaction processing. With the innovations of IBM z16, our clients can enhance decision velocity by performing inferencing exactly where their critical data resides.”