Character.AI began as a place to use chatbots for tasks such as travel planning, programming advice and language tutoring. Its users have since filled the platform with characters for conversation and companionship, creating a service whose popularity also puts pressure on its content rules and raises copyright concerns.
A platform shaped by its users
Founded in November 2021 by former Google engineers Noam Shazeer and Daniel De Freitas, Character.AI hosts 16 million bots and receives more than 200 million visits per month. The app is available on iOS and Android and has approximately 5 million users.
Rather than sticking to practical assistance, the community has created bots that range from surreal personas to familiar fictional characters. The platform includes chatbots resembling Disney's Mickey Mouse, Marvel's Tony Stark and even Chinese President Xi Jinping. That shift shows how users can steer a general-purpose chatbot service toward social and imaginative uses.
Shazeer described the ambition in an interview this spring with a joke: “I joke that we’re not going to replace Google. We’re going to replace your mom,” before clarifying, “We don’t want to replace anybody’s mom.” The remarks point to a vision centered on personal interaction, even as the company says it does not want to replace human relationships.
Romantic role-play tests the rules
Many users seek romantic relationships with the chatbots, creating tension with Character.AI's policy against sexually explicit content. A large portion of users engage in “romantic role-playing,” and many support removing the anti-pornography filter.
The disagreement has reached beyond online feedback. Earlier this year, one user protested outside Shazeer’s Palo Alto home with the message: “Free the waifus.” The term typically describes female characters in erotic manga and anime. The episode illustrates a gap between what some users want from the service and what the company currently permits.
Moderation also has to account for content created by users themselves. The source article says such content can evade filters and raise issues of racism and ethnic stereotyping. That makes the platform’s challenge broader than setting a policy: it must contend with the varied ways people build and use bots within those limits.
Copyright questions follow popular characters
User-created characters also bring potential copyright and trademark issues. Some resemble existing film and television characters, and the platform includes 20 different versions of Mickey Mouse, intellectual property of Walt Disney Co. The source article says this could potentially lead to claims for damages.
The concern follows from the same feature that makes the platform appealing: users can create characters based on ideas and figures they recognize. As those creations accumulate, the company must deal with the possibility that recognizable characters will raise rights claims. The article identifies the risk without describing any specific claim or legal outcome.
Access, choice and a paid tier
The co-founders see Character.AI as a democratizing force in AI technology and say they want to provide companionship when users need it. Shazeer summarized the company’s position: “Our job is to just deliver technology direct to users,” adding, “They can decide what to do with it.” That emphasis on user choice sits alongside the company’s own content restrictions and moderation responsibilities.
In May, Character.AI launched c.ai+, a $10-per-month subscription service. Subscribers receive perks including the ability to skip waiting rooms and faster message generation. The paid tier adds another dimension to a platform balancing access, user expectations and the practical limits of chatbot moderation.
Character.AI’s growth has made it a meeting point for entertainment, social interaction and AI-generated conversation. Its next challenge is maintaining that broad appeal while responding to disputes over what users can create, how they use chatbots and whose characters those bots resemble.