A BILL to amend and reenact §§ 2.2-3906, 6.2-500, 6.2-501, 6.2-506, 6.2-510, 6.2-513, 36-96.1:1, 36-96.3, 36-96.4, 36-96.8, 36-96.10, and 36-96.16 of the Code of Virginia and to amend the Code of Virginia by adding a section numbered 2.2-3905.2, relating to Virginia Human Rights Act; equal credit opportunities; Virginia Fair Housing Law; nondiscrimination by automated decision systems.
HB999 would amend Virginia’s Equal Credit Opportunity laws to restrict how creditors use automated decision systems, including artificial intelligence and machine-learning tools, when making adverse credit decisions. Under the bill, a creditor could not use an automated decision system to deny, revoke, or otherwise take an adverse action on a credit application unless a natural person reviewed the application and approved the final action. The reviewing person would need authority to approve, modify, or reject the system’s evaluation or recommendation.
The bill also requires creditors to maintain policies and procedures designed to ensure compliance, and it extends responsibility to creditors that rely on third-party vendors that develop, provide, or operate automated decision systems. It clarifies that the measure does not create a new basis for liability, expand remedies, alter existing liability allocations under state or federal law, or require disclosure of proprietary or trade-secret information. The State Corporation Commission would be authorized to adopt conforming regulations and issue guidance, and the act would take effect on January 1, 2027.
HB999 would add a new section to Chapter 5 of Title 6.2 governing credit practices in Virginia and would amend related definitions and regulatory provisions in the state’s Equal Credit Opportunity framework. Its practical effect would be to impose a human-review requirement before adverse credit actions can be taken using automated decision systems, affecting lenders, creditors, and their third-party technology vendors. The bill also directs the State Corporation Commission to align regulations with federal Equal Credit Opportunity Act rules and permits guidance on best practices.
No committee transcript or recorded vote information is available, and the bill was left in the House Appropriations Committee. Based on the text, the bill appears to reflect a cautious, consumer-protection-oriented approach to the use of artificial intelligence in credit decisions, while also attempting to avoid broad new liability or disclosure obligations. The lack of recorded floor action suggests the measure did not advance beyond committee consideration.
The main point of contention is likely the balance between consumer protection and operational flexibility for creditors using automated decision tools. Supporters would likely favor the bill’s requirement for human oversight, especially in light of concerns about algorithmic bias, transparency, and fairness in credit underwriting. Opponents or cautious stakeholders may object to added compliance costs, potential delays in credit processing, and uncertainty about how much meaningful review a natural person must provide, particularly when third-party vendors supply the systems. The bill’s explicit disclaimer against new liability and trade-secret disclosure appears designed to address some of these concerns.