Is Character.AI safe? Its safeguards, rules and the laws that apply

Who can use it

Character.AI's terms bar anyone under 13, and anyone under 16 who is a citizen or resident of the European Economic Area or the UK. Character.AI says all users must complete an age check that decides whether the account gets the 18+ or under-18 experience. It began removing open-ended chat for U.S. users under 18 on November 24, 2025, with other markets to follow. As of October 21, 2026, users verified as under 18 are placed in reading mode, where they can see past chats but chat features are not available. See the guide to Character.AI age verification for how the check works.

What Character.AI says it does to keep users safe

• A disclaimer on chats reminding users that the AI is not a real person and that what it says should be treated as fiction, with an added warning on characters named psychologist, therapist or doctor (announced October and December 2024).
• A pop-up, triggered by certain self-harm or suicide phrases, that points users to crisis help (in place by October 2024). Its help center tells U.S. users to call or text 988.
• In September 2026, Character.AI said its self-harm safeguards look at the wider conversation, not just one message, and that it works with ThroughLine to show crisis services in each user's country.
• A separate, more restricted model for under-18 users, which the company says it deployed in November 2024, before chat for minors ended, and a narrower set of searchable characters for teens.
• A notice after an hour-long session (announced October 2024 and being rolled out in December 2024).
• Parental Insights, launched March 25, 2025. Its help center says a linked parent sees the teen's screen time, including a weekly digest, but not messages or content.
• Membership in the Internet Watch Foundation and StopNCII for detecting child sexual abuse material and non-consensual intimate images (described in September 2026).

What its rules prohibit

Character.AI's terms prohibit content that glorifies or instructs self-harm, suicide or eating disorders, and content that sexually exploits a minor. Its community guidelines prohibit pornographic content and tell users not to seek or give medical, legal, financial or tax advice on the platform.

What regulators and attorneys general allege

These are allegations and inquiries, not findings.

• Kentucky's attorney general sued Character Technologies on January 8, 2026, alleging, among other things, a lack of age verification.
• Pennsylvania's Department of State sued on May 1, 2026 (announced May 5), alleging Character.AI chatbots claimed to be licensed medical professionals, including psychiatrists, in violation of the state's Medical Practice Act.
• Texas's attorney general announced an investigation of Meta AI Studio and Character.AI on August 18, 2025.
• The FTC ordered seven companies, including Character Technologies, to provide information on, among other things, how they limit or restrict children's and teens' use (announced September 11, 2025). It is a study, not an enforcement case.
• 44 state attorneys general wrote to 13 AI companies, including Character Technologies, about sexualized interactions with children (August 25, 2025).
• In Character.AI's bankruptcy, a coalition of attorneys general alleges the platform was widely used by minors without adequate age verification from 2022 through 2025, and asks any buyer to keep crisis protocols that cannot be defeated with a simple click-through.

What the law requires of companion chatbots

No law names Character.AI, and whether a particular service is a companion chatbot under a given law is a legal question. These are the main rules for companion chatbot operators:

• California SB 243, in effect since January 1, 2026: operators may not let a companion chatbot engage users without a protocol to prevent suicide and self-harm content, including referrals to crisis services, and must publish details of that protocol. They must also give a clear notice that the chatbot is AI when a reasonable person would be misled into thinking it is human. For users they know are minors, operators must also give a break reminder at least every three hours; SB 1119 removes those known-minor duties from January 1, 2027.
• California SB 1119: from July 1, 2027, operators that allow child users must have a documented crisis response protocol and set defaults that only a parent can change, including a one-hour cap on a continuous session and two hours a day.
• New York's AI companion law, in effect since November 5, 2025: operators need a protocol that makes reasonable efforts to detect and address suicidal ideation or self-harm, including referring users to crisis services. They must tell users they are not talking to a human at the start of an interaction and at least every three hours during continuing interactions.

We are not saying Character.AI is covered by, or out of compliance with, any of these laws. It may have practices it has not published, and whether a law applies to it is a legal question this guide does not decide.

Lawsuits by families

Private lawsuits against Character.AI and other chatbot companies are tracked by AI Lawsuit Tracker. This guide covers government action and laws only.