Sponsors (2)
Official title: To direct the Secretary of Veterans Affairs to seek to enter into an agreement with a federally funded research and development center for the conduct of an independent evaluation of artificial intelligence systems in use by the Veterans Health Administration, and for other purposes.
Introduced July 22, 2026 by Gregory Francis Murphy · Last progress July 22, 2026
The bill increases safety, clinician oversight, and transparency around VA AI—reducing risks and improving accountability—but may slow rollouts, strain VA resources, and face funding and information-sensitivity trade-offs that limit how quickly fixes are implemented.
Veterans will receive independent safety assessments of VHA AI systems that identify patient-safety risks and recommend corrective actions.
Clinicians will gain clearer safeguards and oversight confirming AI is used only as decision-support and that they retain override authority.
Congress, the GAO, and taxpayers will get unredacted findings and a VA corrective-action plan, increasing transparency and accountability for AI in veterans' care.
Veterans may experience delays in receiving AI-driven improvements to care when deployments are paused pending evaluations and corrective actions.
VHA will face added compliance, evaluation, and reporting burdens that could divert staff time and clinical resources away from patient care.
A prohibition on new appropriations may leave VA without sufficient funds to implement recommended fixes, limiting the law's effectiveness and delaying remediation.
Based on analysis of 2 sections of legislative text.
Requires VA to contract an FFRDC within 90 days to independently evaluate high‑risk or widely deployed clinical AI systems for safety, performance, governance, transparency, and cost.
Requires the Department of Veterans Affairs to contract with a federally funded research and development center (FFRDC) within 90 days to perform an independent evaluation of AI systems used, piloted, or being developed for clinical use in the Veterans Health Administration. The review must prioritize at least five high‑risk or widely deployed systems and assess integration, governance, performance, safety, human oversight, transparency, cost, and clinician/veteran impacts. Sets detailed evaluation topics for FFRDCs including patient-harm risk, safeguards to keep AI as decision‑support (not autonomous care), clinician override mechanisms, veteran-centric design and trust, cost and sustainment impacts, documentation/transparency, and reporting of adverse events. The law creates a short timeline for the VA to obtain the FFRDC agreement and requires comprehensive, independent technical and operational assessments of VHA clinical AI tools.