Treasury Touts AI-Enhanced Fraud Detection
Commentary by Steven Lofchie
The Treasury Department showcased the benefits of its AI-enhanced fraud detection efforts for fiscal year 2024.
In a press release, the Treasury stated that its AI-enhanced fraud efforts over the past fiscal year have increased both the amounts recovered and the amount of fraudulent money transactions prevented. The Treasury said that over $4 billion more was recovered in FY2024 than in FY2023. The Department attributed the recovery to the use of improved machine learning ("ML") technologies and data sharing and "payment integrity solutions" program partnerships with federal and state agencies.
The Treasury touted the benefits of its efforts to:
- expand "risk-based screening," which resulted in "$500 million in prevention;"
- identify and prioritize "high-risk transactions," which resulted in "$2.5 billion in prevention;"
- expedite the "identification of Treasury check fraud with machine learning AI," which resulted in "$1 billion in recovery;" and
- implement "efficiencies in payment processing," which "result[ed] in $180 million in prevention."
Commentary
It is ironic that the US Government touts the benefits of AI for use by the Government, while setting an impossible standard (full explainability) for regulated institutions that seek to use AI.