Character.AI

Character.AI is a consumer chatbot service that lets users create their own AI characters and hold open-ended conversations with them in text and voice. Launched publicly in 2022, it became one of the most widely used AI companion applications in the world and one of the first to restrict young teenagers from its core chat experience after scrutiny over the effects of parasocial AI relationships.^[1]

History

Character.AI was founded in 2021 by Noam Shazeer and Daniel De Freitas, former Google engineers who had worked on the LaMDA conversational model and left after the company declined to release a public chatbot. The platform opened to the public in September 2022 and grew rapidly, attracting a user base that skewed heavily toward teenagers and young adults. A funding round in 2023 valued the company at roughly US$1 billion.^[2]

In August 2024 the founders returned to Google under a reported US$2.7 billion licensing arrangement for the company's models, and Karandeep Anand, previously of Meta, became chief executive. The change of leadership coincided with intense legal and regulatory pressure: in October 2024 a Florida family filed a wrongful-death lawsuit alleging that the service contributed to the suicide of a 14-year-old user, and in 2025 the company faced further suits as well as an investigation by the Texas attorney general into allegedly deceptive marketing of AI chatbots to children.

Platform and usage

Characters on Character.AI are defined by their creators using a name, a personality description, an opening greeting and example dialogue, after which the platform's fine-tuned language models role-play the persona. Users can browse and chat with millions of community-made characters — fictional figures, historical personae, tutors and companions — and can interact by text, voice calls and generated images. The product is free to use with an optional subscription, and engagement is driven by long, continuous conversations rather than single question-and-answer exchanges, a design that has been both the source of its popularity and the subject of its safety debate.

Safety and regulation

On 29 October 2025 Character.AI announced that users under 18 would lose access to open-ended chat from 25 November 2025, phasing down through a two-hour daily limit in the interim. Minors were moved to structured, pre-created scenes, videos, stories and streams, which the company argued were easier to make safe because the domain of the experience is bounded. Enforcement relies on an in-house age-assurance model combined with third-party verification for users who dispute a classification. The company described the change as more conservative than its peers and said it would hurt engagement, and separately established a non-profit called the AI Safety Lab to work on safety for AI entertainment features.^[3]

The episode made Character.AI a reference point in the wider regulation of AI companions. In the United States, senators from both parties introduced a bill in October 2025 to bar AI companion products from minors and require chatbots to disclose their non-human status, while California enacted legislation requiring safety guardrails on AI chatbots. Analysts note that the core issue is structural: content filters alone were not considered sufficient to make an unbounded, open-ended conversation safe for children.

>See Also

• AI Companions
• Chatbot
• AI Safety
• PDPA Malaysia
• AI Regulation in Malaysia