On 4 February 2025, the High-Risk Artificial Intelligence Developer and Deployer Act, including testing requirements, was passed by Virginia's House of Delegates. The Act would outline the responsibilities of developers and deployers of AI. In particular, the Act would require developers to conduct pre-deployment testing, including bias detection, performance evaluation, and red-teaming. It would also require developers to update their disclosures within 90 days of any substantial modification to the AI system. Deployers would need to conduct ongoing impact assessments and post-deployment monitoring to track system performance and algorithmic discrimination. In addition, AI systems would be required to comply with recognised risk management frameworks such as NIST AI Risk Management or ISO/IEC 42001. The bill would exempt certain technologies and activities, including anti-fraud systems and research, and would place enforcement authority with the Attorney General, with penalties for violations ranging from USD 1,000 to USD 10,000.
Original source