PYMNTS· PYMNTS·· 16 hours agoAI score58
JPMorganChase’s LPM improves fraud detection by up to 85.5%
JPMorgan Builds One AI Model to Catch 85% More Fraud
AI brief
AI-assisted source summary · not independently verified. Content standards
JPMorganChase has put its Large Payment Model (LPM) into production for fraud detection, authorization prediction, and payment retry prediction. Researchers say that after reading up to six months of transaction history for the same party, the model’s card fraud detection performance improved by 85.5% compared with single-task models; it processes millions of transactions each day, with 99% of requests completed within 60 milliseconds.
Source:PYMNTS · pymnts.com