Virginia's Joint Legislative Audit and Review Commission (JLARC) has found significant gaps in how the state oversees artificial intelligence use by agencies and employees, according to VPM News on October 15, 2026.
The report concludes that the Virginia Information Technologies Agency (VITA), which serves 65 state agencies, has AI standards that have not kept pace with the technology's risks. JLARC surveyed 80 executive branch and independent agencies and issued 14 recommendations. Roughly one third of agencies are using or testing AI tools, including a Department of Corrections system that helps screen bags brought into prisons.
Auditors said VITA applies the same requirements to every AI project regardless of risk, and that agencies may face up to 20 separate reviews before a project advances. Nearly half of agencies using AI called unclear requirements a major or moderate constraint. VITA recently required annual audits of high-risk AI systems but has not defined what counts as high-risk, and its enterprise architecture standard for AI was last updated in November 2023.
About 76 percent of employees at supported agencies use general-purpose AI tools, yet only 28 percent of agencies had written AI-use policies as of June 2026, which JLARC described as a higher-than-necessary risk level.
State chief information officer Mike Watson said VITA agrees with the findings. Recommendations include risk-based standards, a published map of the procurement review process, and mandatory annual AI literacy training for state employees.
Source: VPM News -- https://www.vpm.org/news/2026-10-15/jlarc-ai-report-vita-greer-dickinson-watson/