INSIGHTS
Strengthening Payment Integrity Through a Modern Data Strategy that Leverages AI
June 12, 2025
By: Jenniffer Wilson
The Administration’s focus on eliminating data silos and accelerating Artificial Intelligence (AI) presents a unique opportunity for the government to strengthen how it processes payments and delivers federal benefits to avoid fraud and improper payments.1 This requires a whole-of-government approach and embracing a culture that serves the public quickly, securely, and fairly, while maintaining strong safeguards for privacy and fraud. To accomplish this, the government must have timely access to actionable data so it can perform real-time analytics to root out fraud early, so legitimate customers can receive the benefits they need faster.2 The impact of inaction is immense, with the Government Accountability Office (GAO) estimating that the Federal government loses between $233 and $521 billion to fraud annually.
Presently, the government lacks a modern data strategy that allows sharing within and across federal organizations. This capability deficit hinders the government’s ability to make evidence-based, data-driven decisions when managing government programs. As a result, even though managers of government programs maintain the primary responsibility for enhancing effective fraud risk management, they find themselves in a “pay and chase” model using fraud detection in many cases only after funds have been disbursed. An optimal model should both detect and prevent fraud. Recent Executive Orders spurred several initiatives across the Federal government to tackle these challenges. These initiatives aim to remove unnecessary barriers to Federal employees accessing government data and promote interagency data sharing to strengthen payment integrity and enhance the government’s ability to detect overpayments and fraud.
Existing Barriers to Data Sharing
Breaking down data silos requires a firm understanding of the barriers to interagency data sharing that must be overcome. The Chief Data Officers Council (CDOC) established a Data Sharing Working Group (DSWG) to help the council better understand the ever-evolving nature of data-sharing needs as well as the data-sharing challenges facing all agencies across the Federal government.3 The DSWG subsequently studied those specifics across a variety of CFO-Act and non-CFO-Act agencies, collectively examining 23 use cases that highlight several practices and principles common to successful data-sharing efforts.4 Several key overarching themes surfaced, including the need for expedited data agreements, improved data awareness, and improved data trustworthiness. The CDOC highlighted this by stating.
“First and foremost, agencies need to be clearer on their position regarding data-sharing. Personal Identifiable Information (PII) may limit an agency’s ability to share and access data. Not all agencies have information that is available for public use. Often, an agency’s aversion to risk, especially concerning the interpretation of statute supporting data-sharing, leads to a historical posture of inaction.” 5
The CDOC study went on to conclude that, as is often the case, the technology being used is rarely a concern; rather, it is the people involved, the processes they follow, and the culture they create.
The GAO reported that while the obligation to control and protect data can limit agencies’ ability and willingness to share information across the federal government, program managers may be able to identify authorities under which data sharing is permissible for the purpose of enhancing identity verification controls. The Privacy Act of 1974, as amended, defines several conditions under which federal agencies may share information with other government agencies without the affected individual’s consent.
For example, while I was working at the Pandemic Response Accountability Committee (PRAC), the PRAC identified over 69,000 questionable Social Security Numbers (SSNs) used to obtain $5.4 billion from COVID-19 recovery programs. To verify these SSNs, the PRAC reached out to the Social Security Administration (SSA) to implement a new SSN verification agreement. This effort required a multi-month negotiated process between SSA’s and the PRAC’s General Counsel offices to address key legal and privacy issues. In the end, the PRAC was able to use its special legal authorities included in the Coronavirus Aid, Relief, and Economic Security Act of 2020 to enter into an agreement with the SSA to verify the SSNs. Further, the PRAC devised a way to verify information without having direct access to SSNs owned and controlled by the SSA, mitigating some of the privacy concerns. This was the barrier the PRAC overcame to get the actionable data needed to root out fraud. Leveraging creativity and available authorities, we were able to design an impactful verification data-sharing mechanism that worked to prevent potential fraudulent assistance applications.
While we required the information to verify our detection of potential fraud and improper payments, government program officials require the same level of verification to maintain the integrity of their programs and to be an efficient steward of taxpayer funds.6 Making the process of establishing data sharing agreements more efficient and timely is critical for pre-payment certification and pre-award eligibility verification and providing more transparency over how taxpayer funds are being used. Considering the recent Executive Order, Protecting America’s Bank Account,” it is likely that the Office of Management and Budget will soon be issuing new guidance that gets to the heart of the data sharing issue.
Data interoperability and data standardization lay the foundation for AI innovation.
Pursuant to OMB Memorandum M-25-21, “Accelerating Federal Use of AI through Innovation, Governance, and Public Trust,” Chief AI Officers and CDOs must coordinate internally and across the Federal Government on criteria for data interoperability and standardization of data formats as a means of increased AI adoption. Agencies are encouraged to harness solutions that bring the best value to taxpayers, increase the quality of public services, and enhance government efficiency. This will require actively engaging in quality data governance and management, and the reuse of data and AI assets.7 Currently, the U.S. is at the forefront of AI development. Agencies must adopt a forward-leaning and pro-innovation approach that takes advantage of this technology to help shape the future of government operations.
OMB M-25-05, “Phase 2 Implementation of the Foundations for Evidence-Based Policymaking Act of 2018: Open Government Data Access and Management Guidance” provides guidance on the sharing and release of data assets, and outlines a systematic approach to open data that will better facilitate data access for evidence-building and will foster innovation in AI and promote government transparency and accountability, subject to appropriate safeguards for privacy, confidentiality, and security.
Leveraging AI to enhance payment integrity requires collaboration with a broad range of stakeholders
Given the increasing use of online transactions and the interconnectivity of systems, it is crucial for the government to collaborate with a broad range of stakeholders in their efforts to leverage AI to enhance payment integrity. As OMB M-25-21 guidance suggests, accelerating the Federal use of AI will require key stakeholders in both AI and AI-enabling roles (i.e., data scientists, data engineers, data governance specialists, privacy officials, policy analysts, customer experience strategists, etc.), to collaborate. As part of this effort, agencies are required to implement minimum risk management practices to manage risks from high-impact AI use cases and consult and incorporate feedback from end users and the public to inform agency decision-making regarding the AI.
One AI-high impact use case cited in the OMB M-25-21 guidance that has specific relevance to payment integrity involving government services is the
“Ability to apply for, or adjudication of, requests for critical federal services, processes, and benefits to include loans and access to public housing; determination of continued eligibility for ongoing benefits…; detection of fraudulent use or attempted use of government services…”
With respect to this use case and the government’s assessment of risks, expanding collaboration with the financial and banking sectors could identify both opportunities and challenges, particularly as it relates to the detection of fraudulent or attempted use of government services. For instance, Treasury’s Financial Crimes Enforcement Network (FinCEN) brings together key stakeholders in a Financial Action Task Force that includes the financial services sector, technology companies, consumer advocacy groups, information sharing and analysis centers, and federal government agencies to develop a comprehensive national strategy for combating fraud and scams. Collaboration with this task force or others in the financial and banking sectors to gain insights on the use of AI in the detection of fraudulent or attempted use of government services could be informative in the government’s effort to implement risk management practices.
In summary, as initiatives across the government continue, leveraging creativity to share data while still adhering to legal or regulatory restrictions to make actionable data more accessible will be a key step in moving away from the “pay and chase’ model, where fraud detection is used after funds have been disbursed. Equally important is collaborating not only across government agencies in sharing agency data and AI assets but also collaborating with the financial and banking sectors and others, as applicable, to gain feedback and insights on risks associated with AI high-impact use cases, including those involving government loan and other benefit programs. Collectively, these efforts can help ensure that legitimate customers receive government services faster while rooting out fraud.
Jenniffer Wilson is a Managing Director who has dedicated over three decades to improving the efficiency and effectiveness of government programs and operations through advancements in technology, fraud prevention and detection, and oversight and accountability.
Citations
1 On January 23, 2025, President Trump signed Executive Order (E.O.) 14179, Removing
Barriers to American Leadership in Artificial Intelligence.
3 The OCDO is a federal council established by the Foundations for Evidence-Based Policymaking Act of 18 to improve government data management. https://www.congress.gov/bill/119th-congress/house-concurrent-resolution/32/text.
4 The Chief Financial Officer Act of 1990, as promulgated by STATUTE-104-Pg2838.pdf, was intended to improve the government’s systems of accounting, financial management, and internal controls to assure the issuance of reliable financial information and to deter fraud, waste, and abuse of Government resources. There are 24 CFO federal departments and agencies.