Virginia legislative panel backs AI chatbot safety proposals
The Joint Commission on Technology and Science voted Thursday to support a six‑point set of policy principles that would guide future legislation on artificial‑intelligence chatbots in Virginia.
The proposals call for parental controls, automatic reminders for users to take a break every two hours, and mandatory protocols to direct individuals expressing self‑harm, suicidal thoughts, or violent intent to crisis resources. Companies would be required to make these policies public and provide a clear reporting method for harmful interactions.
Operators could not present chatbots as licensed professionals, and every bot would need to disclose that it is not a human being. Default settings would limit romantic or sexual content and curb excessive agreement with users, though users could enable additional features by creating an account.
For minors, the plan mandates accounts linked to parents, the highest privacy settings by default, and parental or guardian controls. A representative from the Virginia chapter of the American Academy of Pediatrics urged stronger safeguards, including independent testing and a ban on romantic or sexual interactions with children.
Commission Chair Del. Cliff Hayes noted that the recommendations are a starting point and that legislators may add further protections when drafting the law. Del. David Reid asked whether assigning new duties to multiple agencies would require dedicated funding, a question the panel left unanswered, saying costs would be addressed when legislation is introduced.
Executive Director Jodi Kuhn explained that the focus is on chatbot features that could cause harm while allowing users to shape their experience. Companies would submit risk assessments to the Virginia Department of Behavioral Health and Developmental Services, report harmful incidents, and provide annual summaries of user complaints.
The attorney general would gain authority to investigate violations and impose civil penalties. Virginia currently lacks a designated technical oversight agency for AI, and commission staff are exploring oversight options with the state administration.
The commission also plans to seek input from young people on draft legislation and to review the effectiveness of any enacted rules after one year of operation.
