Virginia lawmakers hope to change the way we chat with AI bots: 'Very intriguing and very, very scary'

RICHMOND, Va. — Virginia state lawmakers approved a set of priorities for governing chat bots ahead of the next General Assembly session, with specific concerns about the potential harm posed to children.

In a meeting Thursday, the Joint Commission on Technology and Science approved a set of priorities to guide legislation for next year's General Assembly aimed at governing how chat bots operate in Virginia.

"The realm that we find ourselves in with chatbots and A.I. is both very intriguing and very, very scary all at the same time," Del. Mike Cherry (R-Colonial Heights), a member of the commission, said.

"This is an evolving industry and oftentimes you will have a series of really good actors in the industry and you may have some bad actors," added another commission member, Del. David Reid (D-Loudoun). "And so, one of the things that those six foundational principles are is to be able to establish a baseline to ensure that not only children, but adults are being kept safe using this evolving technology."

The principles are mitigate extreme harms for all users, target design features, mandate transparency and accountability, enable user choice and autonomy, enable legislative learning, and engage affected communities.

"The first is mitigating extreme harms for all users. This is essentially the floor. Regardless of what else legislation does, we think it needs to require chat bot operators to have real crisis protocols in place," JCOTS' executive director Jodi Kuhn said of the first principle. "If a user shows signs of self-harm, suicidal ideation, or violence, the chat bot needs to refer them to actual crisis resources and that policy needs to be published somewhere accessible, not buried in terms of service."

The vote followed presentations over the summer from staff detailing how half of American adults use chat bots — and even more teens do — along with the emotional and mental health risks tied to their use.

"By default, chatbots tend to prolong conversations, and by doing so, there are opportunities for benefits such as those listed on the slide. However, alternatively, the same design features can introduce harms such as emotional dependency," Kuhn said.

Recommended priorities also include requiring transparency and data from chatbot companies, which Reid said can help fine-tune laws as they are implemented in the real world.

"Put into legislation metrics that will help to inform future reporting and we also have to ensure that we're putting into the legislation proper controls to make sure that there's not future abuse," Reid said.

Lawmakers discussed how Virginia had largely held off on legislating artificial intelligence last session because of an executive order from President Trump last year that threatened state broadband funding if it did. Staff said that has not seemed to come to fruition, and other states have moved forward with their own legislation — which Cherry said Virginia can use to its benefit.

"They haven't been doing it very long, so we don't have a lot of data to say it's working or not working. But at least we're not starting from scratch," Cherry said.

While Thursday's meeting focused on governing chat bots, another oversight group is set to receive a presentation next week on recommendations for how Virginia should regulate artificial intelligence as a whole.

CBS 6 is committed to sharing community voices on this important topic. Email your thoughts to the CBS 6 Newsroom.