The U.S. Treasury Department is just the latest federal agency to embrace advances in machine learning and artificial intelligence (AI). The department announced earlier this month that it will use the data-driven technology to prevent fraud and improper payments.
This has already helped prevent and recover more than $4 billion in fraud and improper payments this fiscal year (FY) (October 2023 – September 2024), up from $652.7 million in FY23.
“This increase reflects the dedicated efforts of the Treasury Department’s Office of Payment Integrity (OPI), within the Bureau of the Fiscal Service, to enhance its fraud prevention capabilities and expand its offering to new and existing customers. to expand,” the Ministry of Finance said in a statement. .
Online payment fraud is expected to exceed $362 billion by 2028, according to data from Juniper Research.
AI fights fraud
There are concerns that AI tools, especially generative AI, could help criminals engage in bank fraud and other financial-based scams. However, the Treasury Department is now using the same technology – which can analyze large amounts of data – to detect fraud patterns often used by criminals.
The agency, which did not go into specific details, did outline some of the key ways the technology is already being used:
- Expanding risk-based screening, resulting in $500 million in prevention.
- Identifying and prioritizing high-risk transactions, resulting in $2.5 billion in prevention.
- Accelerating the identification of Treasury check fraud with machine learning AI, resulting in a $1 billion recovery.
- Implementing efficiencies in the payment processing schedule, resulting in $180 million in prevention.
“The Treasury Department takes seriously our responsibility to serve as effective stewards of taxpayer dollars. Ensuring that agencies pay the right person, in the right amount, at the right time is central to our efforts,” said Deputy Secretary of the Treasury Wally Adeyemo. “We have made significant progress over the past year in preventing more than $4 billion in fraudulent and improper payments. We will continue to work with others across the federal government to equip them with the necessary tools, data and expertise they need to stop improper payments and fraud.”
In addition to protecting online payments, the Treasury Department has established and strengthened partnerships with new and high-risk programs to increase access to and use of its payment integrity solutions. That included working closely with federally funded, state-administered programs. In May of this year, the Treasury Department and the Department of Labor announced a data sharing partnership that could provide state unemployment agencies with access to “Do Not Pay Working System” data resources and services through the Unemployment Insurance Integrity Data Hub .
Data-driven protection
Exactly how the agency will protect the data appears to be a closely guarded secret, but Dr. Jim Purtilo, associate professor of computer science at the University of Maryland, shared some thoughts with ClearanceJobs.
“(Treasury) was not specific about what specific techniques were employed for screening, but it appears the agency is looking at a lot more data, and that alone certainly has the potential to fill in the pieces of the puzzle and complete more pictures of fraud, ” he explained.
“As with everything in this industry, data quality will be critical,” Purtilo added. “What is the accuracy of these forecasting methods? False positives? False negatives?”
Although the Treasury Department has not said as much, it is possible that the same tools could be used on tax returns to detect possible tax fraud.
“Good for them for finding more cheats, but if the cost of doing so rises out of proportion to the input or if more people are trapped by a high false negative rate, then I doubt voters will conclude it’s a win ”, Purtilo suggested.
AI isn’t going anywhere and could be a good sign that some departments are being proactive in adopting it.
“Certainly, many agencies and companies can achieve the same kind of quality benefits by using data in smarter ways, but we need to keep a close eye on what independent controls are used to track efficacy,” Purtilo said. “Many AI techniques defy a clear explanation for a given outcome, and we wouldn’t want machines accusing humans of things that bureaucrats simply don’t understand. The stronger predictive algorithms must be coupled with stronger accountability.”
