VA now using 367 AI systems in healthcare and benefits
The Department of Veterans Affairs has disclosed it is now operating 367 artificial intelligence systems across its healthcare, benefits processing, and service delivery operations, according to the newly released 2025 VA AI Inventory. This represents what VA officials describe as steady growth in AI deployment, with health and medical applications making up the largest category of use cases, followed by systems supporting benefits administration, information technology, service delivery, financial management, and administrative operations.
The inventory catalogs AI tools ranging from clinical documentation systems used during medical appointments to automated benefits processing applications and customer service chatbots. One notable system, Ambient AI Scribe, listens to conversations between clinicians and patients during appointments and automatically generates clinical notes, reducing the administrative burden on providers so they can focus more attention on veterans during care visits. The inventory also reveals that more than 50,000 VA personnel are using commercial AI products including Microsoft Copilot Chat, Microsoft Teams Premium, Grammarly, and GitHub Copilot for tasks such as meeting transcription, document drafting, report summarization, and information retrieval.
These AI systems affect all veterans who interact with VA healthcare facilities or file disability claims, though the impact varies depending on which services a veteran uses. Veterans visiting VA medical centers may encounter AI-assisted imaging analysis, diagnostic workflows, disease risk identification tools, or automated clinical documentation during their appointments. Those filing or managing disability claims are increasingly interacting with AI systems that process forms, summarize records, and assist claims processors with documentation.
One high-impact system specifically supporting disability claims processing is TERA Memorandum Automation, which helps claims processors complete toxic exposure memoranda by pre-populating answers using information already in veterans' records. VA reports this system achieved a 98.12 percent accuracy rate and was used to assist with more than 181,000 forms, saving an estimated 54,581 work hours since deployment. The VA.gov chatbot, classified as another high-impact use case, helps veterans locate information, complete tasks, and access support services through self-service interactions.
For disabled veterans, the expansion of AI across VA operations carries both potential benefits and legitimate concerns about accuracy and oversight. The Veterans of Foreign Wars has emphasized that AI should function as a tool rather than replacing human judgment, particularly in processes affecting veterans' benefits and healthcare. Chris Macinkowicz, deputy director of the VFW National Veterans Service, stressed the importance of human review after AI systems generate outputs and the need for VA to continuously verify accuracy months and years after deployment, not just during initial testing.
VA requires all high-impact AI systems to complete an AI Impact Assessment and Risk Mitigation Plan before deployment and says these systems are continuously monitored. According to VA Press Secretary Quinn Slaven, any system that cannot be brought into compliance with governance requirements is discontinued. The inventory shows some previously reported AI systems are now classified as retired, though VA has not specified why individual systems were deactivated. For several healthcare systems, the inventory notes that inaccurate outputs could contribute to downstream clinical or operational impacts if not detected and corrected, while stating that providers receive training and oversight to mitigate those risks.
Veterans who believe an AI system contributed to an error in their care or benefits processing should document the issue and contact their Veterans Service Organization representative. As AI becomes more embedded in VA operations, disabled veterans should remain engaged with their VSOs and stay informed about how these systems affect their specific benefits and healthcare services, particularly when filing claims or appealing decisions where automated processing may have played a role.