From NPR.

A new report finds that more than a third of teenagers using AI chatbots use them for personal companionship, raising concerns about developmental and emotional risks.

The report was released by Aura, an online safety company. Psychologist Scott Kollins, Aura’s chief medical officer, says 37% of conversations between teens and their chatbot companions involve violence: "It is roleplay that is interaction about harming somebody else — physically hurting them."
Parents should keep a close eye on how their kids are using chatbots, says pediatrician Dr. Jason Nagata, at UC, San Francisco: "Parents don’t need to be AI experts. They just need to be curious about their children’s lives and ask them about what kind of technology they’re using and why."

And to tell their teens explicitly that chatbots come with risks.