GAO: VA Disability Benefits Modernization, AI Pose Accuracy Risks—What VA Must Fix Next
GAO says VA’s disability-claims tech modernization could speed decisions, but oversight gaps—plus generative-AI transparency/accountability risks—could undercut accuracy.
A newly issued GAO testimony and report put fresh scrutiny on how the Department of Veterans Affairs (VA) is modernizing disability-claims technology—and how any move toward generative AI could affect accuracy and accountability if VA’s oversight systems lag behind the software upgrades.
GAO’s core message is that modernization isn’t just an IT project. It depends on how VA monitors quality, manages training, and keeps decision-making accountable when workflows become more automated.
What GAO says VA is trying to do
GAO says VA is working to modernize its IT systems for disability compensation claims to increase both efficiency and accuracy. GAO also notes that the Honoring our PACT Act of 2022 required VA to develop a plan aimed at increasing the speed and accuracy of claims processing decisions.
GAO also describes VA’s longer-running effort to move from older systems to electronic platforms—highlighting that past technology modernization has faced recurring management and measurement problems.
GAO’s accuracy warning: oversight gaps can travel with the technology
GAO points to two recurring oversight areas that matter directly to what happens to a claim—moving from medical evidence, to examinations, to final rating decisions.
1) Quality oversight for contracted medical exams
GAO says VA has faced challenges overseeing the quality of exams provided by contracted medical providers (called examiners), including weaknesses in regularly reviewing exams for complex claims and issues tied to incentives and feedback.
2) Training and performance management for claims processors
GAO also says VA has had gaps in how it manages and measures training for claims processors—such as establishing goals, tracking whether required training is completed, and assessing whether training is working.
In GAO’s framing, new tools cannot fix underlying quality-control and training-management failures. If VA’s oversight doesn’t strengthen in step with automation, errors can be processed faster without being caught and corrected.
Why GAO flags generative AI as an accountability risk
GAO is not arguing that generative AI has already produced widespread errors. Instead, GAO warns that generative AI can increase risk and complicate accountability—particularly because it may operate in ways that are difficult to explain, audit, or verify.
GAO also describes a practical accountability problem: if VA and other stakeholders cannot clearly detect errors or determine who is responsible for a decision, it can become harder for VA claims processors—and veterans who may need to appeal—to understand what happened and correct problems.
GAO says VA is exploring AI uses such as further automating parts of the disability-claims process (including document intake and preliminary processing), which GAO says could help—but only with strong governance, transparency, and oversight safeguards.
What Congress is pressing VA to get right
In opening remarks for a House hearing on technology modernization, the chair of the House Veterans’ Affairs Subcommittee on Technology Modernization said the focus should be on how VA manages the claims process—not just on the presence of new software. The remarks also emphasized guardrails, including that no veteran benefit claim should be fully decided by AI or automation, and that quality assurance policies should require a “human in the loop.”
What veterans and claimants should watch next
GAO’s “so what” is that faster systems are only valuable if quality monitoring, training oversight, and accountability controls keep pace.
As VA works through GAO’s recommendations, veterans, families, and advocates may want to watch for:
- Whether VA strengthens and institutionalizes review of the quality of contracted medical exams—especially for complex claims—rather than treating quality as a one-time check.
- Whether VA improves how it manages claims-processor training (including tracking completion and evaluating whether training is producing better decision-making).
- Whether any generative-AI use in disability-claims workflows comes with clear governance, documentation, and traceability so errors and misuse can be detected and responsibility can be established.
Sources
- U.S. Government Accountability Office (GAO) — GAO-26-109137 (VA Disability Benefits: Opportunities and Challenges to Modernizing Technology and Adopting AI)
- House Committee on Veterans’ Affairs — Technology Modernization hearing document (DocumentID=7944)
Discover more from Interactive News
Subscribe to get the latest posts sent to your email.
