Industry Newsdeep learningfraud detectionvisafalse positives
Visa Improves Fraud Detection With Deep Learning
9.1
Visa's head of predictive fraud intelligence John Munn told PYMNTS in a recent interview that fraudsters increasingly use AI to probe defenses, making static rules ineffective. He said Visa uses deep learning models trained on global payments data to detect subtle behavioral deviations, delivering 15%–20% higher authorization rates while reducing false positives. Continuous, pre-authorization monitoring and model updates are essential, he advised.
Key Points
- 1Deploys deep learning models achieving 15–20% higher authorization rates than legacy systems
- 2Highlights fraudsters’ use of AI to probe defenses and exploit static rules at scale
- 3Advises practitioners to monitor pre-authorization events and continuously update models to reduce false positives
Scoring Rationale
High practical impact from Visa's official deep-learning results; slightly limited novelty beyond known industry trends.
Sources
Public references used for this report.
Practice with real Payments data
90 SQL & Python problems · 15 industry datasets
Used by DS/ML engineers at top companies
250 free problems · No credit card
See all Payments problems
