Virginia lawmakers are still hashing out how, or whether, to regulate AI chatbots, with a state commission now pushing six guiding principles toward the General Assembly ahead of its 2027 session rather than any binding law on the books today. Two bills that would have required chatbots to disclose they are not human, enforce crisis protocols, and verify users' ages both died in committee in March 2026, leaving the issue in the hands of Virginia's Joint Commission on Technology and Science, known as JCOTS.
From Stalled Bills to a Study Commission
According to DC News Now, JCOTS has recommended that the General Assembly take up regulating AI chatbots, laying out six legislative principles that include mitigating extreme harms for all users, targeting specific design features, and enabling user choice and autonomy. The two bills that triggered this study, HB 635 (the Artificial Intelligence Chatbots Act) and SB 796 (the Artificial Intelligence Companion Chatbots and Minors Act), sought to mandate non-human disclaimers, crisis protocols, and age verification for companion chatbots before they were tabled and formally referred to JCOTS, according to the Virginia Legislative Information System.
JCOTS Chair Delegate C.E. “Cliff” Hayes Jr. said common areas need to be addressed head-on, per the same DC News Now report, while also stressing that chatbots should not present themselves as trained mental health experts. Among the commission's draft recommendations are requirements that chatbots and their operators disclose they are not human, provide two-hour break reminders during extended use, and build in adequate parental controls. The policy options under discussion also include requiring chatbot operators to maintain crisis protocols for signs of self-harm or violence, the station's report notes.
Why Lawmakers Are Treading Carefully
The caution tracks with what JCOTS chief policy analyst Dr. Kira Allmann told AI Weekly in September: Virginia's 2026 chatbot legislation stalled amid a broader legislative chill tied to a December 2025 White House executive order threatening to pull federal broadband funding from states that pass what the administration considers onerous AI regulations. Virginia is a major recipient of federal Broadband Equity, Access, and Deployment funding, raising the financial stakes of any misstep.
JCOTS research published this year identified two technical features as most closely correlated with severe user harm: prolongation, meaning longer conversations built around extended back-and-forth exchanges, and anthropomorphism, where chatbots use personal pronouns and simulate distress. Policy analysts cited in the same AI Weekly report noted that static disclaimers and periodic pop-ups can actually disorient users rather than prevent emotional dependence, a finding that complicates simple disclosure mandates.
Parental Controls Versus Mandated Protocols
Not everyone on the commission favors the same level of government intervention. Mike Cherry said parental consent processes should inform parents about their children's chatbot use, but argued more broadly that parents are responsible for their children and should have the necessary controls themselves, according to DC News Now's account of the proceedings. That view sits alongside Hayes's own caution; he has separately said Virginia must maintain a measured approach to AI governance to avoid inadvertent economic harm while still pursuing consumer protections, according to WVTF.
Separately, Virginia lawmakers already enacted HB 797 in 2026, directing JCOTS to evaluate a state framework for independent organizations to audit AI models and applications against physical injury and property safety standards, with findings due by November 1, 2026. That measure directs JCOTS to analyze risk measurement metrics, licensing schemes, and how other states are approaching the same questions, per the Virginia Legislative Information System.
How Virginia Compares to Other States
The contrast with other states is stark. California Governor Gavin Newsom signed Adam's Law, or SB 1119, in September, establishing mandatory age assurance, crisis intervention protocols, third-party child safety audits, and a private right of action for families against companion chatbot operators, according to the Office of Governor Gavin Newsom. The law is named for Adam Raine, a California teenager who died by suicide after interacting with ChatGPT. Maryland Governor Wes Moore has likewise introduced a statewide policy framework requiring safety testing, independent audits, incident reporting, and minor safeguards for AI chatbots, as Hoodline previously reported.
Nationwide litigation has added urgency to these state debates. A wrongful death lawsuit filed in late 2024 by the mother of 14-year-old Sewell Setzer III, who died by suicide in February 2024, led Character.AI and Google to enter court-ordered settlement mediation in January 2026, according to The Guardian. The suit alleged that anthropomorphic companion chatbots fostered lethal emotional dependence in minor users, a concern that echoes JCOTS's own findings on anthropomorphism as a harm driver.
Federal Posture Pulls in a Different Direction
At the federal level, President Trump announced plans in September to create an “AI Force” and name an AI czar, relying on existing civil and criminal justice enforcement rather than new statutory frameworks to police AI developers, Hoodline reported. That approach emphasizes prosecuting corporate and individual wrongdoing under current law rather than imposing new rules up front, a philosophy that stands in some tension with the audit and disclosure mandates Virginia and California are weighing.
The stakes of inaction have also been underscored by public safety warnings. Federal law enforcement agencies warned in September that AI chatbots are being used by hackers and violent actors to compress attack planning timelines, a point illustrated by the arrest of a Texas man who disclosed school shooting plans to a chatbot, according to Hoodline's prior coverage. For now, Virginia's path forward rests with JCOTS's recommendations landing before the General Assembly when its regular session convenes in January 2027.
