A BILL to amend and reenact §§ 9.1-101, as it is currently effective and as it shall become effective, and 9.1-102 of the Code of Virginia and to amend the Code of Virginia by adding in Article 1 of Chapter 17 of Title 15.2 a section numbered 15.2-1723.3 and by adding in Chapter 1 of Title 52 a section numbered 52-11.7, relating to Department of Criminal Justice Services; law-enforcement agencies and sheriff's departments; policy on use of artificial intelligence systems.
HB1170 would require Virginia law-enforcement agencies, sheriff’s departments, and the Department of State Police to adopt written policies governing the use of “covered” artificial intelligence systems in criminal investigations. The bill defines covered AI broadly to include tools used to generate investigative leads, write or materially aid police reports, and technologies such as biometric identification, facial or location tracking, automated license plate readers, predictive policing, risk scoring, behavioral analysis, fraud detection, social media analysis, and other data-analytics tools. It also excludes ordinary administrative software and generative AI used only for spelling or grammar checks.
The bill directs the Department of Criminal Justice Services (DCJS) to create a model policy by October 1, 2026, and requires each local law-enforcement agency, sheriff’s department, and the State Police to adopt a policy that meets or exceeds that model by January 1, 2027. The required policies must be publicly available on agency websites and must address authorized uses, data collection and retention, sharing practices, and limits on discriminatory or unauthorized uses. The model policy must also acknowledge that violations may lead to administrative discipline.
The bill amends DCJS’s statutory duties in § 9.1-102 to add responsibility for developing a statewide model policy on covered AI systems used by law-enforcement agencies and sheriff’s departments. It creates new Code sections in Title 15.2 and Title 52 requiring local law-enforcement agencies, sheriff’s departments, and the Department of State Police to adopt written AI-use policies consistent with the DCJS model. In practical terms, the bill would impose statewide policy standards on investigative AI use, increase transparency through public posting, and likely affect procurement, training, recordkeeping, and oversight practices for agencies using AI-enabled investigative tools.
The available voting history suggests the bill faced substantial resistance in committee: it was tabled in the House Communications, Technology and Innovation Committee by a 21-1 vote. That outcome indicates limited support for advancing the measure in its introduced form. No committee transcript is provided, so the record does not show detailed debate, but the broad tabling vote suggests concern or reluctance among committee members about the bill’s scope or policy implications.
The main points of contention are likely the breadth of the AI definition and the degree of regulation imposed on law enforcement. The bill reaches a wide range of investigative technologies, including predictive policing, biometric identification, license plate readers, social media analysis, and data integration platforms, which could raise concerns about operational flexibility, privacy, and surveillance. Supporters would likely emphasize transparency, accountability, and limits on discriminatory use, while opponents may object that the bill could constrain investigative tools or require agencies to adopt restrictive policies before statewide standards are fully developed.