Criminals are leveraging artificial intelligence to create sophisticated fraud schemes that outpace the federal government's ability to detect and prevent them, a fraud expert told Congress this week. David Maimon, head of Fraud Insights at SentiLink, testified before the House Oversight Committee's Government Operations Subcommittee, warning that the government lacks the necessary tools, policies, and collaboration to combat the evolving threat.
"What's happening right now is the fact that we simply don't have the right tools to deal with fraud," Maimon told Fox News Digital. "We don't have the right policies to deal with fraud. We don't have enough deterrence. We have no close collaboration between the government and the [private] sector to try and find the right solution to this issue."
AI-Driven Fraud Tactics
Maimon highlighted how criminals are using AI to generate fake documents, videos, and phishing emails. These tools can be used to bypass identity verification systems, including liveness checks that use selfies and video, and to steal personal information. "It's difficult to prove that you're not essentially using AI while verifying your identity with documents and liveness tests and selfie images," he explained.
During his testimony, Maimon noted that fraudsters are already employing AI-generated faces and deepfake videos to defeat digital identity checks used by banks and tax preparers. He shared insights from his company's investigations, which have uncovered online marketplaces where criminals trade stolen identities, fraudulent checks, and tutorials on how to commit fraud against government programs and financial institutions.
Exploiting Systemic Weaknesses
Maimon also pointed out that criminals exploit vulnerabilities between government agencies. "Fraud against government programs is no longer a series of isolated schemes," he stated. "It is a durable, specialized criminal infrastructure." He explained that while government defenses are often built on a program-by-program basis, criminals' infrastructure is designed to operate across multiple systems.
He cited an investigation into Medicaid fraud where his team found providers billing the government millions of dollars, claiming to have staff and caregivers. However, upon physical inspection of the claimed addresses, many of the employees and operations did not exist. This highlights the need for both digital and physical verification, as well as the reliance on trusted historical data.
The Need for Modernized Defenses
To counter these advanced threats, Maimon suggested that the government should move beyond solely relying on images or videos for identity verification. He proposed incorporating historical identity data, which is more difficult for AI to falsify, into verification processes. "The bad guys are moving faster than the government," Maimon stated. "While the government is still trying to catch up and roll out better tools, criminals keep finding new ways to steal money."
He emphasized that every dollar protected from organized fraud is a dollar that remains available for the intended beneficiaries of government programs. However, Maimon cautioned that current law enforcement efforts have only uncovered a fraction of organized fraud, suggesting they are "just scratching the surface."