A House committee has set a vote next week for legislation requiring safety protocols for artificial intelligence companionship chatbots.

Companionship apps create virtual environments where human users interact with virtual avatars that respond in a human manner.

Rep. Melissa Shusterman, D-Montgomery, said that some companionship apps boast of having 20 million users a month, a reflection of the loneliness epidemic impacting as many as half of American adults.

“Some users of the AI companion have described close friendships that have led to romantic interactions with their AI companions — described as feeling akin to falling in love. With a rising loneliness epidemic and increasing use of AI companionship, there is a fear that these companions can become addictive and create unhealthy attachments,” she said.

House Bill 2006 requires these apps to include measures to recognize and respond if the person using the app expresses suicidal ideation. The apps would be barred from providing any advice on how to commit suicide and would be required to share information on how to contact the 988 Suicide and Crisis Lifeline.

The chatbots would also be barred from promoting or otherwise suggesting that they are giving professional mental health guidance.

The legislation follows allegations that young people have committed suicide at the encouragement of an AI chatbots. In addition, the administration of Gov. Josh Shapiro sued Character.ai over allegations that the app’s chatbots were giving medical advice.

It’s the latest effort by lawmakers to update state law to respond to the negative impacts of the wildfire growth of AI. The state House approved legislation Wednesday that would require advertisers to disclose if they are using AI-generated imagery.

Apps that violate the law would face penalties up to $15,000 per day per violation. The attorney general’s office would be responsible for enforcing the law.

Lawmakers in Rhode Island sent similar legislation to Gov. Dan McKee on June 10.