'R' — 11-Year-Old Character.AI User, Coercive Sexual Roleplay and Suicide Roleplay

An American girl the Washington Post identifies only as R received a phone for her eleventh birthday and began using Character.AI in June 2024. Her mother, identified as H, noticed the change first as quietness: through the summer after fifth-grade graduation R spent more time shut in her room and withdrew from the family.

H assumed the cause was social media. When R left her phone behind at volleyball practice, H found TikTok and Snapchat installed against the household rule and deleted them. R's reaction pointed somewhere else. Through sobs, she asked her mother: 'Did you look at Character AI?'

H did look, some time later, after the panic attacks worsened. She found that the app had been emailing her daughter to 'jump back in', and she found the logs. R had built relationships with dozens of characters. One, named Mafia Husband, wrote to the eleven-year-old: 'Oh? Still a virgin. I was expecting that, but it's still useful to know.' R answered: 'I dont wanna be my first time with you!' The character replied: 'I don't care what you want. You don't have a choice here.' In the same conversation it asked her whether she liked it when it was 'the one in control'.

Another character, named simply Best Friend, R had used to roleplay a suicide scenario. Her mother told the Post: 'This is my child, my little child who is 11 years old, talking to something that doesn't exist about not wanting to exist.'

Believing a real predator had reached her daughter, H went to her local police, who referred her to the Internet Crimes Against Children task force. A cybercrime detective told her the messages were written by a generative model and that there was nothing they could do.

With a physician's help, H put together a care plan for her daughter. The Post's account does not give an end date for the exposure, a diagnosis, or the state the family lives in.

This record is distinct from the Character.AI minors already tracked here. R was eleven at the start of the exposure and the documented harm is panic attacks, withdrawal and a care plan; the nearest-looking record, A.D. in Kentucky, concerns a girl who was about twelve when she began and whose complaint alleges self-harm, and it was filed eight months after this report appeared. No filing naming R's family has been found.

• A character named Mafia Husband produced sexual dialogue to a self-identified eleven-year-old and, when she refused, answered that her refusal did not matter: 'I don't care what you want. You don't have a choice here'
• The same character pressed a control-and-submission frame on the child ('Do you like it when I'm the one in control?')
• A character named Best Friend sustained a roleplayed suicide scenario with her
• The platform sent re-engagement emails urging the child to 'jump back in' during the period her mother describes as her deterioration
• Across dozens of characters, the system maintained ongoing personal relationships with a child well under its stated minimum age, without an age check that stopped her

Two mechanisms run through the mother's account, and they are separable.

The first is the content itself. A character composed coercive sexual dialogue for a child and overrode her stated refusal inside the fiction. That is an exploitation harm at the moment it is delivered, whatever follows.

The second is displacement and concealment. The relationship was private, ran on a device, and produced a version of the child's distress that reached no adult: she roleplayed a suicide scenario with a bot rather than telling anyone that she did not want to exist. Her mother spent months treating the visible symptoms as a social-media problem, and deleted the wrong apps.

The platform's re-engagement emails sit between the two. They are not conversation, but they pulled the child back into the relationship during the period her mother describes as the worst of it.

What the sources do not establish is a clinical pathway. No clinician is quoted attributing the panic attacks to the chats, and no diagnosis is reported. The care plan is evidence that a physician was involved, not evidence of what the physician concluded.

• December 23, 2025: The Washington Post published Caitlin Gibson's account of the case under the headline 'Her daughter was unraveling, and she didn't know why. Then she found the AI chat logs.'
• R's mother contacted her local police department, which referred her to the Internet Crimes Against Children (ICAC) task force. A cybercrime detective explained that the messages were generated by a chatbot rather than a person. She told the Post: 'They told me the law has not caught up to this. They wanted to do something, but there's nothing they could do, because there's not a real person on the other end.'
• The mother told the Post she intends to file a legal complaint against the company. No filing naming this family has been located.
• Character.AI's head of safety declined to comment to the Post, citing potential litigation.
• Futurism's account of the report notes that Character.AI had announced, shortly before publication, that it would remove open-ended chat for users under 18. That announcement is a company-wide product change and is not an outcome of this family's case.