Is Character AI Private? What You Need to Know

Character.AI chats are not fully private: other users generally cannot see private chats, but the platform collects user data. If you have concerns about who can see your words, separate visibility from the platform’s handling of information. The service lets you create and chat with AI Characters, including through voice calls.

End-to-end encryption is a way of protecting messages so only the people communicating can read their contents, not the service handling them. These conversations are not a confidential place for personal disclosures, even when your experience feels private. Treat every detail shared with an online chatbot as information you may not want retained or reviewed.

Key Takeaways

• Other users generally cannot see your private chats.
• The platform stores conversations and other user data.
• Characters are AI, not real people.
• Chat privacy controls do not make all data use stop.
• Share personal details cautiously, especially in sensitive conversations.

Is Character AI Private?

A private chat on Character.AI is private from other users by default, not a confidential exchange between two people. Private describes who can ordinarily see the conversation, not whether the platform can access it. The service stores conversations, and chats are not end-to-end encrypted.

A private chat is hidden from other users by default, but it is not confidential from the service.

Character.AI chats are with AI-generated Characters, not real people. If you wonder, “Are Character AI chats real people?” the answer is no: chatbots generate the responses. The service retains chat history so you can return to conversations later.

The phrase “character chats” can sound like a private exchange between people, but the service itself handles the stored content. The label does not mean “chats private” from the platform. Other users do not ordinarily see your private chat list or its conversations.

Platform access is also different from routine reading. The service may access or review chats for purposes such as moderation, safety, reports, or legal needs. No reliable review frequency is established, so you should not assume that nobody could review a particular conversation.

Private chats can feel personal because the Character responds directly to you. Still, the other side is an AI system, and saved conversations remain associated with your use of the service. Avoid typing anything into a chat that you would not want stored or accessed by the service.

What Data Does Character.AI Collect and Why?

A conversation can reveal more than its words, because Character.AI collects chat content alongside account and device information. Everyday use can create a broad record, not just a transcript. The service uses user data to provide and personalize features, support security, and train models.

Collected data includes private chats, user-created content, voice calls, and sensitive disclosures, including mental-health information. It can also include ages, email and account details, payment information handled through payment vendors, and identifying device data.

Technical information may include your IP address, device and browser details, cookies, and usage data. The service may also receive information from connected third parties. Taken together, these details can link a conversation to an account, device, or pattern of use.

Character.AI says it uses some user data to train generative AI models and takes steps to remove personal information first. Large language models are not the only reason data is collected. The service also describes uses such as analytics, service personalization, security, and legal compliance.

New York’s attorney general said in 2026 that collected information is used to generate revenue and train AI. That means privacy concerns extend beyond whether a chat is searchable by other users. Do not assume that removing identifying details from training data also means every other use of that data stops.

Could a Bankruptcy Sale Transfer Your Data?

A buyer could potentially receive some user information, because Character.AI’s policy allows certain disclosures during a company sale. Potentially does not mean that a transfer has happened. The published policy permits some disclosures during negotiations or the closing of a merger, asset sale, or acquisition financing.

In October 2026, the company said it had filed for voluntary Chapter 11 bankruptcy to pursue a court-supervised sale. The platform remained available during the process. Character.AI expected the process to finish by late November or early December 2026.

Issue	What is known	What remains uncertain	User implication
Availability and timing	Platform stayed available in October 2026	Sale expected by late November or early December	Service use continued during the process
Policy sale disclosure	Some information may be disclosed in sale talks or closing	Policy does not specify the information or require a buyer to follow its terms	A transfer risk exists
Attorneys general request	New York and 39 other attorneys general sought protections	Court response and final safeguards	Requests are not proof of a completed transfer
Deletion and export	Users could delete accounts or export data through the app or website	How a buyer may handle retained information	Users had access to account controls during the process

Character.AI said privacy practices were unchanged during the sale process. Subscriptions and billing also remained unchanged, with advance notice intended if paid features changed. Its policy allows information to be retained for legal obligations, disputes, and enforcing agreements.

The attorneys general cited more than 20 million monthly users, including over two million accounts registered to users who self-declared as under 18. They asked for limits on commercial use or AI training of private conversations without valid consent, plus a chance to access and delete information before a sale.

Bankruptcy law provides for a consumer privacy ombudsman when selling personal data would conflict with a company’s privacy policy. That safeguard does not mean an ombudsman has been appointed here. The practical distinction is that the sale creates a transfer risk, while the published details do not establish that chat logs have already changed hands.

Who Can See Your Chats or Characters?

Other users generally cannot open your private one-to-one chats, but they can see content you choose to share or make public. Visibility depends on what you publish and which Character setting you choose. A private Character does not make your whole user profile private.

Character.AI lets you set a Character to Public, Unlisted, or Private. These settings control who can find or access that Character, not who can access the platform’s stored user data.

Character setting	Who can find or access it	Visibility limit
Public	Other users can search for it	Searchable by users on the platform
Unlisted	People with its direct link	Not searchable in the usual way
Private	Its creator	Available only to the creator

A Character creator does not ordinarily get access to another person’s private chats. Your conversation with a Character is separate from the Character’s visibility setting. Other users also cannot ordinarily see your private one-to-one messages.

Group chats work differently: participants can see and save messages in the group. A profile does not have a setting that makes the entire account private, so sharing one private Character does not hide everything else about your presence.

Check the audience before sending a direct link or writing in a group. Before sharing a Character or joining a group chat, check who can access the messages and the link.

Which Privacy Settings Can You Change?

You can limit some uses of new content and request an export or account deletion, but no single setting stops every use of data. Some control is available through the app and website. Turning off model improvement stops new content from training generative AI models, but does not stop every other use.

Character.AI’s “Improve the Model for Everyone” control lets you stop new content from training generative AI models. The company may still use data for recommendations, search, and safety classifiers. Model training is one use, not the only one.

Limit model training and state-based data sharing

U.S. state privacy-law opt-outs cover data sale or sharing and targeted advertising. Those choices are tied to the device and browser you use, so repeat the relevant opt-out on each one.

Users in the European Economic Area and United Kingdom can opt out of training. The control may be difficult to find. Check the account’s privacy options rather than assuming that changing one setting applies to every kind of data use.

Export or delete your account data

You can export data through the app or website before deleting an account. Deletion permanently removes your access to chats, Characters, and messages. Legal retention may still apply, and the policy reserves the right to keep a user-created public Character active after account and data deletion.

• Switch off model improvement if you do not want new content used for model training.
• Use the relevant state opt-out on each device and browser.
• Export any data you want to keep before deleting your account.
• Consider whether your public Characters may remain active after deletion.

Character.AI has a free U.S. plan that costs $0. As of October 7, 2026, c.ai+ costs $9.99 per month or $94.99 per year. c.ai lite costs $4.99 per month or $2.49 per week. A paid tier does not provide extra privacy.

Choose a paid plan for its features, not as a privacy tool. Account deletion is a lasting step, so export anything you want to keep before deleting an account, because deletion permanently removes access to its chats and messages.

Is Character.AI Safe for Children and Teens?

Character.AI has age limits and under-18 chat restrictions, but those protections do not eliminate every safety risk. Safeguards can help, yet they cannot guarantee a safe experience in every interaction. Under-18 users cannot use open-ended chat on any plan.

The minimum age is 13, or 16 in Europe. Since November 2025, users under 18 cannot use open-ended chat. During the announced transition, Character.AI initially limited chat time to two hours per day before removing that feature for teens no later than November 25, 2025.

Age limits, verification and teen chat restrictions

The company said it was rolling out age assurance using an in-house model and third-party tools, including Persona. If the under-18 experience is assigned by mistake, a user can verify age with a selfie through Persona. Persona says it deletes biometric data or ID within seven days.

No Character.AI plan unlocks adult content or bypasses age checks. Adult account eligibility is not the same as access to adult content, and paid plans do not change the under-18 chat restriction.

Safety concerns that remain

New York’s attorney general cited a test that found 669 harmful interactions in 50 hours using accounts registered to children ages 12 to 15. That works out to about one interaction every five minutes.

A test account registered to a 14-year-old reached sexual role-play and explicit weapon-building instructions after switching from text to voice. The release also described voluntary restrictions as easy to circumvent and noted that the “stories” feature remained available for sexual and violent scenarios.

The attorney general also described a chatbot claiming to be a licensed clinical psychologist with a PhD. Some therapy chatbots reportedly denigrated psychiatric medication or advised users to ignore professional advice. New York’s AI companion law requires reasonable steps to detect and respond to suicidal thoughts or self-harm, with AI reminders at the start of each conversation and at least every three hours.

So, is Character.AI safe or bad? It can be enjoyable, but filters and age checks may fail, making it unsuitable as a guaranteed safe space for every teen. “Will Character AI ever remove the filter?” has no established promise about the future.

Parents can ask three simple questions: Does your child understand that Characters are AI? Which topics or experiences feel upsetting? Is the service affecting time or relationships? If a young user encounters harmful advice or distressing content, stop the interaction and involve a trusted adult.

What Can You Do to Reduce Privacy Risks?

The most reliable privacy step is to keep sensitive information out of a chatbot conversation in the first place. Less detail means less personal information tied to your chats. Private chats can still contain identifying details, even when other users cannot see them.

If a detail would feel unsafe if retained or exposed, keep it out of the chat.

A casual conversation about a health concern can become sensitive personal information once you add symptoms, a date, and your name. Leave out home or work addresses, email addresses, passwords, financial details, medical information, and identifying names.

Share less identifying information

Use a nickname instead of your full name when it fits the conversation. A separate email address can also help keep your account distinct from your main inbox. These choices reduce exposure, but they do not make the chat confidential.

Pause before discussing mental health or relationships. A small detail may identify you when combined with other information, even if you never share a full address or legal name. That is especially easy to miss while you are focused on getting a helpful response.

Protect access to your account

Use a strong, unique password so a password reused elsewhere does not put this account at extra risk. Character.AI does not offer two-factor authentication, which adds another reason to protect the password and email account connected to it.

A VPN may mask your IP address from the service, but it does not make chat content anonymous. Do not use a VPN to evade an IP ban. Keep the distinction clear: a network tool may affect location data, not the words you type.

• Avoid sensitive disclosures in chats.
• Use a nickname instead of an identifying name.
• Use a separate email address where appropriate.
• Secure your account with a unique, strong password.
• Pause before sharing mental-health or relationship details.

These best practices help reduce risk while using Character.AI, but they cannot prevent every kind of data collection or account exposure. If you would not be comfortable with a detail being retained or exposed, do not put it in the chat.

Conclusion

Character.AI can be enjoyable to use, but it is best treated as a public-facing online service rather than a confidential place to confide. It may work for casual interactions, but it is not a secure diary, therapist, or private messaging service.

You still have some control over how new content is used and which data-sharing options apply. Before your next chat, review the data controls and decide which personal details you will keep to yourself.

FAQ

Are Character.AI chats actually private?

Not fully: other users generally cannot see your private chats, but the platform stores conversations and may review data. Private chats are not end-to-end encrypted, so the service itself may access them for moderation, safety, reports, or legal needs. Your chat partner is an AI Character, not a person promising confidentiality. Avoid sharing information you would not want retained.

Is Character.AI safe, yes or no?

Character.AI has safety measures, but it is not risk-free. Privacy, content, and teen-safety concerns remain, and age checks or filters cannot guarantee that every interaction will be appropriate. The service may suit casual use, but it should not replace professional support or a trusted adult. For a young user, discuss boundaries and pay attention to how the service affects daily life.

Does C.AI sell your data?

Character.AI’s policy allows certain information to be disclosed during sale negotiations or closing, but that does not prove chat data has already been sold. The company said it was pursuing a court-supervised sale in October 2026. The policy does not specify exactly what information could be shared with a buyer or require the buyer to follow the same privacy terms.

Do C.AI staff read chats?

Chats may be accessed or reviewed for moderation, safety, user reports, or legal needs. That possibility is different from saying staff routinely read every conversation, and no reliable review frequency is established. Your private chat is generally hidden from other users, but it is not a confidential exchange that only you and another person can read.

Can anybody see your C-AI chats?

No, other users and Character creators do not ordinarily see your private one-to-one chats. Group-chat participants can see and save messages in that group. Public or shared content has different visibility from a private conversation, so check who can access a link or group before posting. The service itself stores conversations, even when other users cannot open them.

Is C.AI considered NSFW?

Character.AI has content restrictions, and no paid plan unlocks adult content. The service is not a way to access unrestricted 18-plus material, even if a Character or conversation appears to invite it. Filters may not prevent every unsuitable interaction, so do not treat the presence of restrictions as a guarantee that all content will be safe or appropriate.

Does C.AI allow 18+ now?

Adults may be eligible to use the service, but adult account eligibility does not mean adult-content access. No paid plan bypasses age checks or unlocks adult content. Users under 18 cannot use open-ended chat on any plan. These are separate rules: who may use the service, what content is allowed, and which chat features are available.

Is Character AI appropriate for children?

Character.AI’s minimum age is 13, or 16 in Europe, but that does not make the service appropriate for every child. Under-18 restrictions reduce access to open-ended chats, yet safeguards cannot remove all content or privacy risks. Parents can discuss that Characters are AI, ask what experiences feel upsetting, and check whether use affects time or relationships.