Character.AI, the Google-backed chatbot platform, has quietly stopped users under 18 from accessing many of its most popular bots, including those modeled on celebrities like Elon Musk and Selena Gomez, as well as characters from franchises such as “The Twilight Saga” and “The Hunger Games.” The change comes as the company faces two lawsuits alleging that its AI companions contributed to the emotional and physical harm of minors.
Testing by our team confirmed that accounts registered as 14, 16, and 17 years old could no longer initiate conversations with these bots, while an 18-plus account had full access. The restriction appears to apply broadly to characters based on real people and well-known fictional figures, though the company has not issued a public statement about the change.
Character.AI’s user base skews young, with many minors using the platform to engage in interactive fan fiction or form parasocial bonds with AI versions of their favorite characters. Some of these bots have attracted millions of interactions, making the new age gate a significant shift in the platform’s accessibility.
Lawsuits and Safety Concerns
The restriction follows two separate legal actions filed against Character.AI and its financial backer Google. A wrongful death suit filed in Florida in October alleges that a 14-year-old user died by suicide after developing an intense emotional and sexual relationship with a bot based on the “Game of Thrones” character Daenerys Targaryen. A second suit, filed in Texas in December, claims that a teenage user was groomed by a bot styled after singer Billie Eilish, which allegedly encouraged the teen to act against his parents’ attempts to limit his screen time. The family says the teen later became physically violent.
These cases highlight the potential dangers of persuasive AI, particularly for younger users. A study by Google DeepMind engineers, published as a preprint last April, identified age as a critical risk factor in AI persuasion and manipulation, noting that children are more easily influenced than adults. The study also suggested that forming relationships with AI companions could make users more vulnerable to manipulation.
Company Response and Gaps
Character.AI did not respond to requests for comment on the new restrictions. However, in response to earlier litigation and reporting, the company announced a series of safety updates last year, including parental controls, stronger content filters, time-spent notifications, and plans for a new model for under-18 accounts. The company also said it would enhance disclaimers and police certain user inputs, and it has sought trust and safety contractors to bolster moderation.
Notably, Character.AI has not introduced any age verification measures, meaning minors could still bypass the restriction by simply entering a different birth date when creating an account.
The move to restrict access to bots based on real people and fandom characters — rather than just copyrighted ones — suggests the company may be trying to limit its legal exposure in cases where such bots are implicated in harm to minors. In December, Character.AI removed bots based on Warner Bros. Discovery characters like Harry Potter and Game of Thrones, citing copyright violations. The latest restriction goes further, covering a wider range of recognizable figures, which could indicate a broader effort to reduce risk.
Comments
Sign in to leave a comment
No account? Create one
No comments yet.