Anti-fraud and anti-money laundering companies are looking to AI to increase the efficiency of their detection models as digital fraud becomes more pervasive. 

Account takeovers using stolen credentials and related financial crime increased by 354% year over year in the second quarter, according to the September “Q3 2023 digital trust and safety index” report by Sift, a fraud-fighting solution company.

“Fraudsters are getting better and sophisticated with their attacks digitally,” Armen Najarian, chief marketing officer at Sift, told Bank Automation News. “To have a good customer experience and increase customer lifetime, a good fraud-fighting mechanism has been more essential.” 

Increasing fraud 

According to the Sift report: 

• Global fraud losses are expected to be 20% higher than last year; 
• Account takeovers increased 808% among fintechs in Q2 YoY;  
• Total account takeover-related losses are expected to exceed $635 billion in 2023 compared to $343 billion in 2022; and 
• 67% of account takeovers resulted in funds being drained from accounts. 

The rise of generative AI is creating new challenges in distinguishing synthetic IDs, deepfakes and social engineering scams, like text or email phishing, Najarian said. 

“We know that the attackers are using generative AI to produce synthetic identities to execute bot attacks, to obfuscate their identities,” which has prompted Sift to experiment with gen AI across multiple scenarios to fight fraud, he said. 

In fact, use of fraudulent synthetic identities has increased 17% in the past two years in the United States, according to a report published today by Deduce, an AI-driven identity verification company. 

“AI is enabling attackers to do two key things: carry out more sophisticated attacks and carry out attacks that are greater scale and velocity,” Gartner analyst Akif Khan told BAN. “The defenders are very much in a proverbial arms race with the attackers.” 

Gen AI to fight fraud 

Fraudsters are using gen AI to “create phishing content such as emails or SMS messages that are much more convincing and devoid of the usual spelling mistakes,” and “using automated bots to carry out large-scale credential stuffing attacks,” Khan said. 

Account takeover remains a critical challenge, and attackers are increasingly targeting the account recovery or password reset processes, he said.  

“After all, why bother stealing a password when you can just reset one?” Khan said. 

Attacks on biometric processes is another area with increased activity, he added. 

While fraud is increasing in quantity and sophistication, banks are upping fraud fighting efforts, Khan said. 

“Most security vendors have been using machine learning, which is a form of AI, for many years now and continue to evolve,” Khan said. “Being able to distinguish between a human or a bot, for example, involves having to examine thousands of signals from an online banking session and make a quick and accurate decision.” 

Citibank and Citizens Bank, for example, work with Akamai Technologies, an AI-driven security company, to stop phishing and bot attacks, and Wells Fargo is using Arkosa Labs’ AI capabilities to fight fraud. 

“At Citizens, we are actively leveraging AI and analytics approaches to address fraud through a combination of real-time data and advanced modeling techniques,” Beth Johnson, chief experience officer at Citizens Bank, told BAN. “To enable these outcomes, we have invested in our data foundations, ensured we have the right technology toolkits and also established a strong risk framework.” 

Get ready for the Bank Automation Summit U.S. 2024 in Nashville on March 18-19! Discover the latest advancements in AI and automation in banking. Register now.