VA Claims AI Accountability Act
- Sponsor
- Rep. Barrett, Tom [R-MI-7]
- Committees
- Veterans' Affairs Committee (primary)
- Last action
- Sep 23, 2026
Bottom line
This bill aims to significantly upgrade the VA's disability claims system using advanced technology and AI, while embedding safeguards for human oversight, data control, and accountability to improve veteran experience and claims accuracy.
What it actually does
This bill mandates the Secretary of Veterans Affairs to establish a program to design, build, and operate a technology platform to modernize the disability claims process, leveraging artificial intelligence. It requires human review for all claims analyses and final decisions, ensures the VA retains unlimited data rights, and mandates data interoperability across relevant agencies. The bill also establishes an AI governance framework and requires extensive congressional oversight and an independent evaluation by the Government Accountability Office.
Proponents argue
Supporters argue this bill is a critical step towards streamlining the often-cumbersome disability claims process, reducing wait times, and improving accuracy for veterans through modern technology and AI. They emphasize that the mandated human review and robust data governance provisions ensure accountability and protect veterans' due process rights, preventing fully automated decisions.
Opponents contend
Potential opponents might express concerns about the ambitious timeline for implementing such a complex IT modernization and AI integration project, fearing potential cost overruns, delays, or disruptions to existing services. They might also raise questions about the inherent risks of relying on AI for sensitive claims decisions, even with human oversight, citing potential for algorithmic bias or errors.
The bill is concise and clearly structured, allowing for a thorough understanding of its provisions within a reasonable reading time.
Section 2(a)-(c)
Establishment of VA Disability Claims Modernization Program
This section requires the Secretary of Veterans Affairs to create a program to design, build, implement, and operate a technology platform. The platform's purpose is to modernize the disability claims process, focusing on improving veteran experience, increasing efficiency and accuracy, enhancing system architecture and data infrastructure, and improving employee efficiency and user experience.
Supporters argue
Supporters argue this program is essential for addressing the long-standing challenges and backlogs within the VA's disability claims system, ultimately delivering better and more timely outcomes for veterans and optimizing federal resources.
Critics contend
Critics might contend that large-scale government IT modernization projects frequently encounter significant delays and cost overruns, potentially disrupting current services before delivering the promised benefits.
Tradeoffs
The bill navigates the tension between the urgent need for technological advancement in VA services and the inherent risks and complexities associated with implementing major government IT projects.
Section 2(d)
Mandated Human Review and Final Decision-Making
This provision explicitly requires that any automation or artificial intelligence used in claims analysis must include human review of individual conditions claimed for benefits. Furthermore, it mandates that employees of the Department of Veterans Affairs retain responsibility for making all final decisions regarding the approval of claims and the granting of benefits.
Section 2(e)-(g)
Data Control, Interoperability, and Security Standards
The bill mandates that the Department of Veterans Affairs retains 'unlimited rights' in all technical data, including source datasets and AI outputs, generated by the new platform. It also requires the platform to establish direct, automated data interoperability between the Veterans Health Administration, Veterans Benefits Administration, and the Department of Defense. Additionally, the technology platform must achieve a 'high impact' categorization under the Federal Risk and Authorization Program (FedRAMP).
Section 3
Artificial Intelligence Governance Framework
This section requires the Secretary of Veterans Affairs to establish a comprehensive framework for governing the use of artificial intelligence within the new technology platform. The framework must advance the program's objectives, ensure that AI use does not infringe upon veterans' due process rights, and clearly differentiate between standard automation and AI. The Department's Chief Artificial Intelligence Officer is designated as the primary official responsible for approving and enforcing this framework.
Sections 4 and 5
Congressional Oversight and Independent Evaluation
The bill mandates significant oversight, requiring the Secretary to submit an annual report to Congress detailing a life-cycle cost estimate, program activities, performance assessment, and average confidence scores of AI agents. Quarterly briefings are also required. Furthermore, the Comptroller General of the United States (GAO) must conduct an independent evaluation of the program's performance metrics and effectiveness, submitting a report to Congress.
Section 6
Program Termination Date
This section stipulates that the program established under Section 2, including the authority to operate the technology platform, shall automatically terminate on October 1, 2028.
The bill defines 'artificial intelligence' by referencing section 5002 of the National Artificial Intelligence Initiative Act of 2020 (15 U.S.C. 9401).
Section 3(b)(3)
Why it matters:It is standard legislative practice to avoid re-defining terms that are already established in federal law, promoting consistency across statutes.
Case for: Proponents would argue that this ensures consistency with existing federal definitions of AI, providing clear guidance for agencies already operating under that definition and avoiding ambiguity.
Case against: Critics might suggest that requiring external lookup for a core definition could make the bill less accessible to readers unfamiliar with the referenced statute, potentially obscuring the full scope of the bill's application.
Estimated impact: Clarifies the precise technological scope of the AI governance framework and the types of systems it will cover.
The program to establish the technology platform is directed to 'leverag[e] existing agreements where practicable.'
Section 2(a)
Why it matters:It is likely intended as a practical instruction for efficient implementation rather than a central policy point, hence its placement within the operational details.
Case for: Proponents would argue this promotes efficiency and cost-effectiveness by encouraging the utilization of existing contracts and infrastructure, thereby avoiding redundant procurement processes and accelerating implementation.
Case against: Critics might suggest that an over-reliance on existing agreements could limit the VA's ability to adopt the most innovative or tailored solutions, especially if current contracts are not optimally suited for advanced AI development.
Estimated impact: Could significantly influence the speed, cost, and technological flexibility of the program's implementation strategy.
The technology platform must be categorized under the Federal Risk and Authorization Program (FedRAMP) at the 'high impact' categorization level.
Section 2(g)
Why it matters:It is a technical standard, often placed in specific requirements sections rather than the main policy directives.
Case for: Proponents would argue that this ensures the highest level of security for sensitive veteran data, protecting against breaches and ensuring the integrity and resilience of the system.
Case against: Critics might point out that achieving 'high impact' FedRAMP categorization is a rigorous, time-consuming, and often costly process, which could add significant delays and expenses to the project, potentially challenging the 2028 termination deadline.
Estimated impact: Imposes stringent security and compliance requirements, significantly affecting the project's timeline, budget, and choice of technology vendors.