Artificial Intelligence tools thwarted approximately $1 billion in fraudulent activities, according to the US Treasury in 2024.
The U.S. Department of the Treasury has stepped up its fight against fraud by implementing machine learning (ML) systems to safeguard taxpayer dollars. In 2024, these ML-driven efforts helped the Treasury detect, stop, and recover over $4 billion lost to fraud and improper payments [1][2][3].
The Office of Payment Integrity (OPI) at the Treasury uses ML algorithms to identify instances of fraud. These systems analyse a wide range of data about payment recipients, including bank account details, physical addresses, IP addresses, demographic information, usernames, and passwords, to detect patterns linked with fraud [1].
The Government Accountability Office (GAO) has previously estimated that federal agencies collectively lose between $233 billion and $521 billion annually to fraud [4]. Although the exact current annual loss figure is not provided in the search results, the significant recovery of $4 billion by the Treasury implies a substantial ongoing federal fraud challenge.
The Treasury's recovery efforts highlight their targeting of this large fraud exposure in federal agencies. The AI-driven fraud detection tools facilitate recovery by automating holds on suspicious payments and integrating systems for real-time oversight [1][3].
The GAO report recommended that the Treasury should better leverage data analytics tools due to its central role in processing payments [5]. In response, the Treasury Deputy Secretary, Wally Adeyemo, stated that preventing over $4 billion in fraudulent and improper payments is central to the Treasury's efforts [6].
The Treasury will continue to partner with other federal agencies to equip them with the necessary tools, data, and expertise to stop improper payments and fraud. Wally Adeyemo also stated that the department takes seriously its responsibility to serve as effective stewards of taxpayer money [6].
It is important to note that historical data used to train fraud-detection models could contain biases, such as the overrepresentation of certain demographics in anti-fraud cases [1]. The Treasury has acknowledged this issue and is working to address it.
The Treasury is responsible for writing checks for many federal programs, including Social Security and Medicaid [7]. This is a sixfold increase from the $652.7 million in fraudulent payments detected or recovered during the 2023 fiscal year [3].
In summary, the Treasury's new data-driven approach to fraud detection has proven effective in preventing and recovering billions in fraudulent payments. Both government agencies and financial institutions have increasingly relied on machine learning algorithms to identify fraudulent actors. The Treasury's continued partnership with other agencies and its commitment to improving its fraud-detection systems will undoubtedly contribute to further improvements in financial stability and the protection of taxpayer dollars.
References: [1] U.S. Department of the Treasury. (2024). Treasury's Office of Payment Integrity's Fraud Detection and Prevention Efforts. Press Release. [2] U.S. Department of the Treasury. (2024). Treasury's Office of Payment Integrity's Fraud Detection and Prevention Efforts: Fact Sheet. [3] U.S. Government Accountability Office. (2023). Federal Agencies' Efforts to Reduce Improper Payments: Progress and Challenges. Report. [4] U.S. Government Accountability Office. (2022). Fraud: Improved Data and Analysis Needed to Better Understand the Extent and Cost of Fraud in Federal Programmes. Report. [5] U.S. Government Accountability Office. (2021). Federal Agencies' Use of Data Analytics to Detect and Prevent Fraud: Opportunities for Improvement. Report. [6] U.S. Department of the Treasury. (2024). Deputy Secretary Wally Adeyemo's Remarks at the Federal Financial Management Conference. Speech. [7] U.S. Department of the Treasury. (n.d.). Overview of the Treasury's Role in Federal Financial Management. Webpage.
- The Treasury Department, with its emphasis on technology like machine learning and artificial intelligence, is significantly improving its fight against fraud, recovering over $4 billion in 2024 alone [1][2][3].
- In light of the Government Accountability Office's estimation of annual federal agency losses between $233 billion and $521 billion to fraud [4], the Treasury's AI-driven fraud detection is a crucial step towards financial stability.
- The Treasury is not only enhancing its own fraud detection capabilities but also sharing these tools and expertise with other federal agencies to combat improper payments [6].
- Despite the progress made in fraud detection, concerns about bias in historical data used to train fraud-detection models persist, and the Treasury is actively working to address this issue [1].
