A new VA Office of Inspector General report examines the Veterans Health Administration’s use of generative AI chat tools, including VA GPT and Microsoft 365 Copilot Chat, for clinical care and documentation. The review found widespread staff engagement with these tools, particularly for drafting clinical notes and summarizing patient information.
The OIG’s primary concern centers on oversight and patient safety. While VA has designated certain AI-enabled clinical tools as high-impact and subject to formal safeguards, the report found that similar risk management practices have not been applied to general-purpose AI chat tools used in clinical settings. The review also identified limited coordination with patient safety programs and the absence of AI-specific mechanisms for identifying and tracking potential safety events.
For government contractors, the report highlights the federal government’s growing focus on AI governance, risk management, human oversight, quality assurance, and patient safety as agencies expand the use of generative AI. VA concurred with the report’s recommendations and outlined plans to develop additional guidance, oversight processes, and safety monitoring capabilities.
Readers should review the full report for details on the findings, recommendations, and planned corrective actions.
