Artificial Intelligence contributed significantly to federal authorities in uncovering over a billion dollars worth of fraud within a year, with its capabilities continuing to expand.
Machine learning AI significantly aided the U.S. Treasury Department in navigating large datasets, leading to the recovery of around $1 billion in check fraud during fiscal 2024, as per recently disclosed figures. This figure is nearly triple the amount recovered in the preceding fiscal year.
Renata Miskell, a high-ranking Treasury official, shared this information with CNN, emphasizing that using data has significantly boosted the department's abilities in detecting and preventing fraud.
The Treasury Department credits AI for playing a pivotal role in preventing and recovering over $4 billion worth of fraud during fiscal 2024, a sixfold increase compared to the previous year.
The utilization of AI in detecting financial nefarious activities began discreetly towards the end of 2022, with the U.S. officials drawing inspiration from the methods employed by numerous banks and credit card companies to quell illicit activities.
Fraudulent activities escalated during the Covid-19 pandemic as the U.S. federal government rapidly disseminated emergency aid to consumers and enterprises, leading to an increase in such incidents.
However, the Treasury Department has not incorporated generative AI, which has gained widespread popularity through platforms like OpenAI's ChatGPT and Google's Gemini, capable of crafting music and answering intricate queries, albeit with occasional challenges in handling basic queries.
Instead, the Department relies on machine learning, a subset of AI with proficiency in scrutinizing extensive data and making educated decisions and predictions.
Machine learning, with its ability to analyze immense volumes of data at a significantly faster pace than humans, aids in identifying subtle patterns and potentially harmful transactions. Once sophisticated AI models are trained, they can swiftly identify suspicious financial activities.
"Fraudsters are extremely deceptive, aiming to discreetly manipulate the system," Miskell mentioned. "AI and the systematic utilization of data enable us to track down buried patterns and anomalies, and make efforts to thwart them."
Given the Treasury Department's status as one of the largest payers globally, it is a critical target for fraudsters looking to capitalize on taxpayer funds.
The Treasury Department handles about 1.4 billion annual payments worth almost $7 trillion for over 100 million individuals. It manages payments such as Social Security, Medicaid benefits, federal worker paychecks, tax refunds, and stimulus check distributions.
Thus, the Treasury Department is a prime target for numerous fraudsters.
Last year, the Internal Revenue Service (IRS) announced the deployment of AI to flag tax evasion by scrutinizing intricate and elaborate returns submitted by hedge funds, law firms, and others.
By 2028, online payment fraud is projected to surpass $362 billion, as per Juniper Research estimates.
Moreover, AI itself is being used to amplify some of these fraudulent activities.
In a notable case earlier this year, Hong Kong authorities claimed a finance worker was swindled out of $25 million after being deceived by a deepfake video.
U.S. officials have raised concerns regarding the threats posed by AI to the financial sector. In June, Treasury Secretary Janet Yellen warned bankers of the "significant risks" associated with AI in the financial sector.
Top regulators, including Yellen, classified AI as an "emerging vulnerability" for the financial system at the end of 2022.
Miskell emphasized that AI systems will alert authorities to suspicious transactions but that final fraud determinations are made by humans.
The Treasury Department's use of AI in combating financial crime is a nascent development.
Miskell hinted that the Treasury Department is investigating ways to incorporate advanced fraud detection methods utilized by leading banks and credit card companies, without disclosing further details to avoid aiding potential fraudsters.
A Treasury spokesperson informed CNN that the department is accelerating its efforts to enhance the tools available for detecting fraud in federal- and state-administered programs. Officials are exploring new data sources to improve fraud detection and collaborating with state agencies to combat unemployment insurance fraud.
The Treasury Department's reliance on machine learning in detecting fraud has led to a substantial increase in recovered funds, with around $1.5 billion being recovered through AI in fiscal 2025, expanding upon the previous year's achievements.
Businesses within the financial sector, such as banks and credit card companies, have been instrumental in inspiring the Treasury Department's adoption of AI technologies for fraud detection, contributing to a more secure financial landscape.
