Enhancing Government Payment Integrity: Leveraging AI and Other Emerging Technologies

The event brought together federal agency leaders, congressional staff, industry, and academic experts to discuss the complexities of improving payment integrity. Participants explored challenges, opportunities, and solutions for using artificial intelligence (AI) to tackle improper payments, which includes fraud, waste, abuse that could affect overpayments, underpayments, and disbursements to recipients.

The roundtable's insights form the basis of a new report from the IBM Center, Enhancing Government Payment Integrity: Leveraging AI and Other Emerging Technologies, by Richard Hoehne of Georgia State University and Karen Kunz of West Virginia University.  The report which highlights specific issues and proposes next steps for enhancing government payment integrity using AI and related emerging technologies.

The report first highlights several challenges faced by government in reducing improper payments, including data sharing restrictions, insufficient investment in analytics and AI, the ineffectiveness of the "pay and chase" model, and a lack of consistent and measurable incentives to reduce errors. Roundtable participants expressed concerns about keeping up with evolving technology used by adversaries, balancing identity protection with customer experience, and dealing with agency siloes and outdated laws.

The authors next identify three specific challenges that emerged from the discussion: data complexity, technology limitations, and user experience. The complexity and variability of data across multiple agencies and programs present a significant hurdle to applying AI for reducing improper payments. Each agency operates with its own data standards, formats, and systems, making comprehensive integration and analysis challenging. Additionally, outdated legal frameworks, such as the Privacy Act of 1974, hinder modern data sharing protocols that can ensure strong privacy and security without having to address unnecessary compliance requirements. The volume and diversity of data, along with potential biases in algorithms, further complicate efforts to identify and stop improper payments.

Balancing robust identity verification with user experience is another challenge, as agencies must ensure secure login and authentication without creating unnecessary delays for legitimate users. Overly stringent security measures for access can deter eligible individuals from receiving benefits, especially vulnerable populations. Addressing these challenges requires not only technological upgrades but also advancements in skills, measures, and organizational culture to adapt effective practices.

The report also identifies three main areas of focus to improve payment integrity in government: collaboration, data and AI, and technology.   First, collaboration across agencies, industry partners, and international organizations can enhance the identification and mitigation of improper payments. Sharing research, data, and best practices can standardize procedures and policies, reducing inconsistencies that adversaries might exploit.

Agencies can focus on modernizing data and AI strategies, standardizing common data structures, and defining ubiquitous controls, while balancing privacy and payment integrity protection. AI can significantly enhance the accuracy and efficiency of identifying risky transactions and shifting from a "pay and chase" approach to pre-payment detection.

The technology opportunity emphasizes layered defenses, combining advanced AI-driven behavior monitoring with a broader view of data to detect fraud patterns. Monitoring policies for gaps exploitable by adversaries can enhance data sharing and counter-fraud measures. Opportunities for improvement include accountability and a skilled workforce, and the report proposes. 9-part counter fraud framework. Implementing these strategies can significantly enhance payment integrity and reduce improper payments.

The report concludes with a summary or recommended steps that emerged from the roundtable for agencies to reduce improper payments, improve efficiency, and enhance customer experience. Key recommendations include:

• Enhancing data sharing protocols through standardized procedures and systems, modernizing the Privacy Act of 1974, and developing a data/algorithm clearinghouse.
• Leveraging predictive analytics to help agencies uncover hidden trends and anomalies, enabling proactive fraud detection and enhancing payment process efficiency.
• A centralized fraud detection hub that can facilitate real-time data analysis using advanced AI and machine learning technologies, identifying patterns and anomalies indicative of fraud before payments are made.
• Investing in advanced fraud detection tools.
• Utilizing machine learning for data consistency.
• Enhancing identity verification and monitoring to further improve payment integrity.
• Regular training and awareness programs for staff on the latest fraud detection techniques and AI tools.
• Establishing clear governance and accountability structures, including designating an accountable senior official and defining roles and responsibilities.
• Collaborating with industry experts to help agencies stay updated on the latest fraud prevention technologies and practices.