NEW YORK — Mastercard's generative AI fraud detection now identifies stolen credit and debit card numbers at double the previous rate — catching compromised credentials before criminals can exploit them, the payments giant announced in May 2024.
The upgrade to Mastercard's anti-fraud system uses artificial intelligence to analyze patterns in stolen card data, scanning billions of cards and millions of merchants to spot signs of compromise. Banks are now notified up to 300% faster, giving cardholders the chance to receive replacement cards before fraudsters can act.
How Mastercard's AI Detects Stolen Card Numbers
The system works by identifying partial card numbers posted on illegal dark web marketplaces and using generative AI-based predictive technology to reconstruct the full 16-digit card details. It factors in geographic location, transaction timing, and other contextual signals — examining over 40 security and infrastructure criteria — to determine which cards have likely been compromised.
"With the help of generative AI, it is possible to determine where your credentials might have been stolen, how it could occur, and how it can be fixed — not only for you but for other customers who do not even know that their information has been stolen," said Johan Gerber, EVP of Security and Cyber Innovation at Mastercard.
Key Improvements Over Previous Fraud Detection
The numbers are significant. Mastercard's new AI-driven approach delivers three measurable gains:
- Detection rate doubled for compromised cards - False positives reduced by up to 200% when identifying fraudulent transactions - Speed increased by 300% in identifying at-risk or compromised merchants
The improved pattern recognition can also pinpoint the original source of a breach — such as a hacked merchant or payment processor — by analyzing batches of fraudulent cards. Gerber noted this capability far surpasses what is achievable through conventional database searches or manual analysis.
The Scale of Dark Web Card Fraud
Currently, billions of stolen credit and debit card numbers circulate on the dark web, typically obtained through merchant data breaches or compromised point-of-sale terminals. These stolen credentials can go undetected for years, surfacing only when payment networks discover the data, a merchant reports a breach, or a card is used for an unauthorized transaction.
An estimated $1 billion is lost to card skimming alone each year, with criminals posting stolen payment details online within hours of a breach.
Faster Replacements, Less Disruption
The practical benefit for cardholders is straightforward: faster notification and faster replacement.
"The good news is that we can now contact banks to make certain that those consumers are serviced and receive new cards as soon as possible, thereby disrupting their lives as little as we can," Gerber said.
The payments industry is also in the midst of a longer-term shift away from static card numbers toward transaction-specific identifiers. That transition is expected to take years, particularly in the United States where adoption of newer payment technologies has historically been slow. In the meantime, Mastercard's AI-powered detection provides a critical layer of protection within the existing system.
