The use of AI in fraudulent activities has spanned various schemes, from impersonating voters through phone calls to creating fake celebrity giveaways. In response, the US Treasury Department has highlighted the significant role of machine learning AI in bolstering its fraud detection capabilities over the past year.
In a recent announcement, the Treasury reported successfully preventing and recovering over $4 billion in fraud and improper payments in the last fiscal year, from October 2023 to September 2024. This figure marks a substantial increase from the previous year’s total of $652.7 million. A significant portion, one-fourth of the $4 billion, was recovered by accelerating the detection of Treasury check fraud using machine learning AI.
Moreover, the efforts included $2.5 billion from identifying and prioritizing high-risk transactions and an additional $680 million allocated towards other prevention strategies. The Treasury intends to share this technology with other federal agencies, although some agencies, like the IRS, have already begun adopting AI to address tax evasion, automate services, and conduct audits.