Fraud detection

The Knoble, a nonprofit that brings banks, credit unions, payment networks and law enforcement together to detect elder and child exploitation and human trafficking, has worked with more than 600 banks.

California has a pending law that would hold banks liable for three times the amount of a loss if they had reason to suspect financial exploitation of an older customer and did nothing about it.

Machine learning has a firm place in banks' fraud and scam detection work. But some are also experimenting with using generative AI to research and handle claims.

Specialized large language models should be harnessed to help financial institutions identify and halt fraudulent activity. The best model would involve regulator-approved pooling of anonymized customer data.

Companies lose an estimated 5% of their revenue each year due to fraud, according to a report from the Association of Certified Fraud Examiners.

When customers are targeted directly with fake text messages that lead to account takeover, artificial intelligence, and in some cases generative AI, can play a role in fighting the fraud, experts say.

Fraudsters are leveraging advanced artificial intelligence to impersonate customers and deceive financial institutions, but the industry is fighting back with detection systems.

One area of focus for the bank is using advanced artificial intelligence to detect business-email compromise. The payment messaging network Swift and online gambling host Caesars are also using AI to stop people from gaming their systems.