AI and VA Disability Claims: A Double-Edged Sword for Veterans
Introduction
AI and VA disability claims are transforming the veteran benefits landscape. While artificial intelligence offers new avenues for veterans to navigate the complex claims process, its implementation raises questions about efficiency and accuracy. This article explores how AI is being used to process claims and the experiences of veterans and former employees.
A Veteran’s High-Tech Approach to Filing Claims
Not all Vietnam veterans are embracing artificial intelligence, but some, like “Chip,” a Marine Corps veteran and former software developer, are finding innovative ways to use it. Chip, who asked to be identified by his nickname for privacy, utilized an AI chatbot to assist with his disability claim for Agent Orange-related cancer and PTSD. He inputted his symptoms and requested the AI to generate a narrative for his supplemental claim. After carefully reviewing the output, he submitted it to the Department of Veterans Affairs. Just a month and a half later, his claim was approved with a 100% disability rating, significantly increasing his monthly benefits.
The VA’s Embrace of Automation
Chip’s experience highlights a broader trend: the VA is increasingly incorporating AI into its disability claims process. The agency has been developing its own AI systems to help claims processors manage the vast amounts of evidence veterans submit. This technological push aims to address a long-standing backlog and expedite response times. The VA has reported a significant reduction in its claims backlog, with processing times dropping from an average of five months to under three months, and in some cases, even faster. VA Secretary Doug Collins stated a goal of processing claims within 30 to 40 days, a previously unprecedented target. The agency attributes this progress to investments in automation technology, aggressive hiring, and the utilization of overtime.
Automated Decision Support: A Million-Dollar Investment
The VA has committed substantial funding to artificial intelligence programs for claims processing. A key initiative, Automated Decision Support, is part of a large contract with IBM. This system is designed to sift through crucial evidence in a veteran’s file, along with federal records from the Department of Defense and the VA, to create a summary document for claims processors. This summary aims to streamline the evidence review, which previously consumed a significant portion of the claims processing time. Paul Shute, former assistant deputy under secretary at the VA, noted that gathering and synthesizing information used to account for about 86% of the time a claim was pending.
AI’s Role in Presumptive Conditions and Beyond
Automated Decision Support was initially deployed for presumptive conditions linked to the PACT Act, such as cancers resulting from toxic exposures. However, its application has since expanded to over 170 VA diagnostic codes. While the system offers the potential for faster processing, its early implementation revealed some challenges. A 2023 report from the VA Office of Inspector General identified issues such as the system failing to recognize duplicate evidence, identifying false evidence, and missing relevant information, which could lead to inaccurate claim decisions. In response, the VA has implemented enhancements to the system, and its progress has been acknowledged.
Concerns About Quality and Accuracy
Despite advancements, some professionals express concerns about the quality of AI-assisted claims processing. Attorneys who litigate veterans’ disability cases report that their staff must meticulously review evidence to ensure the AI hasn’t overlooked or misinterpreted crucial information. Examples include veterans being incorrectly informed they used the wrong forms. Former VA claims raters have described the Automated Decision Support system as “hit or miss,” noting that while it increased processing speed, they often had to correct errors or locate missed information, questioning whether quality has truly improved.
The Pursuit of Speed vs. Accuracy
The VA’s accelerated drive to process claims has sparked debate about whether speed comes at the expense of accuracy. The agency’s own data shows that the three-month accuracy rate for determining individual medical issues correctly has hovered around 94%, below the agency’s target of 98%. The accuracy rate for processing entire claim packets is even lower at 84%. While this accuracy rate is the highest it has been in two years, it has been below target for over a decade. Some veterans’ advocates and legal professionals argue that the focus on speed leads to rushed decisions with potentially significant errors, which can be detrimental to veterans already hesitant to seek benefits.
Investments in New AI Technologies
The VA continues to explore and invest in new artificial intelligence applications. In April, it was noted that the VA has been piloting an AI program since August 2025 that uses veterans’ medical records to complete disability benefit questionnaires. This aims to reduce the need for in-person medical examinations. The agency’s website lists a similar program in its predeployment phase. Additionally, an AI program called Smart Ratings Recommendation, which would propose specific ratings for claims processors to review, was being considered. However, following concerns about potential misuse, this initiative has been indefinitely paused.
The Role of Hiring and Overtime
Beyond AI, the VA has employed traditional methods to combat the claims backlog. The agency significantly increased its hiring of claims processors, adding over 10,000 between 2021 and 2024. However, hiring freezes in subsequent years led to an increased reliance on mandatory overtime for existing staff. This practice, which has been in place intermittently for years, was intensified, requiring processors to work additional hours each month. While some appreciated the extra pay, others found it demoralizing and felt it compromised their work-life balance and potentially the quality of their reviews.
Challenges in the Appeals Process
The increase in claims processing speed has historically correlated with a rise in appeals. When claims are processed more quickly, there is a greater likelihood of denials or errors, which veterans then appeal. This dynamic has been described as a “Whac-a-Mole” scenario. While the overall number of appeals has increased, the VA maintains that the appeal rate has remained steady. For some veterans, like John Cordle, the appeals process, while lengthy for the initial claim, was relatively swift once it reached the Board of Veterans’ Appeals, ultimately leading to a favorable outcome and increased benefits.
Conclusion
The integration of AI into the VA disability claims process presents both opportunities and challenges. While technologies like Automated Decision Support aim to increase efficiency and reduce backlogs, concerns about accuracy and the human element of decision-making persist. The VA’s ongoing efforts to balance speed with quality, coupled with its investments in both AI and traditional staffing, will continue to shape the experience of veterans seeking benefits.
Frequently Asked Questions
What is AI and how is it being used in VA disability claims?
AI refers to artificial intelligence, and in the context of VA disability claims, it’s being used to help process claims by analyzing evidence, summarizing information, and potentially generating narratives or rating recommendations.
Can veterans use AI tools like ChatGPT for their disability claims?
Yes, some veterans, like Chip, have used AI tools to help draft narratives for their claims, which they then carefully review and submit to the VA.
What is Automated Decision Support?
Automated Decision Support is a VA initiative, part of a contract with IBM, that uses AI to sort through evidence in a veteran’s file and compile summaries for claims processors.
What were some of the initial flaws found in the VA’s AI systems?
Early reports indicated that the VA’s AI systems sometimes failed to recognize duplicate evidence, identified false evidence, and missed relevant information.
Has the VA addressed the flaws found in its AI systems?
Yes, the VA has implemented enhancements to its AI systems in response to watchdog findings, and its progress has been acknowledged.
Are there concerns about the accuracy of VA disability claims processed with AI?
Yes, some attorneys and former VA employees express concerns that the drive for speed, partly due to AI, may lead to errors and impact the quality of claim decisions.
What is the current accuracy rate for VA disability claims?
The three-month accuracy rate for determining individual medical issues correctly is around 94%, which is below the VA’s target of 98%. The accuracy for processing entire claim packets is lower.
Besides AI, what other methods has the VA used to reduce its claims backlog?
The VA has also focused on aggressive hiring of claims processors and implementing mandatory overtime.
Has the VA paused any of its AI programs for claims processing?
Yes, a program called Smart Ratings Recommendation, which would have provided proposed ratings for claims processors, has been paused indefinitely.
What is the VA’s goal for processing disability claims?
The VA is working towards processing claims within 30 to 40 days, a significant reduction from previous average processing times.
