New York, Thursday, October 24, 2024-The US Department of the Treasury used machine learning AI to detect and recover more than $4 billion in fraudulent or improper payments in 2024, a dramatic six-fold increase from $652.7 million in 2023.
The Treasury’s Office of Payment Integrity (OPI), part of the Bureau of the Fiscal Service, credits AI for flagging high-risk transactions and speeding up the identification of check fraud.
In a press release, US Deputy Treasury Secretary Wally Adeyemo stressed the importance of AI in safeguarding taxpayer money and ensuring correct payments.
With online payment fraud expected to rise beyond $362 billion by 2028, the Treasury aims to continue partnering with federal agencies to boost fraud prevention.
According to the report, AI models helped prevent $2.5 billion in fraud, and the Treasury’s check fraud systems recovered $1 billion.
Additional steps, such as improved payment processing schedules, blocked $180 million in improper payments, while expanded screening methods averted $500 million in potential fraud.
ELNA.ai founder Arun PM warned that AI systems can be vulnerable if not maintained properly, highlighting risks like overfitting and flawed judgment.
However, Arun emphasized the global importance of fraud prevention and the potential for AI to have a larger impact.
The Treasury’s efforts align with other federal agencies, such as the IRS, which has used machine learning to detect tax evasion, and NOAA, which deployed neural networks to enhance hurricane forecasting in 2024.
