Is Character.AI Safe? Safer for Teens, Not for Their Data

Ask whether Character.AI is safe, and you get two different answers depending on what you mean.

The company removed open-ended chat for everyone under 18 in November 2025, after lawsuits and pressure from regulators. On teen safety, it went further than any of its peers.

On data, it went the other way. Character.AI collects your chats, the images and videos you post, your date of birth, your approximate location, and voice recordings if you use voice features. Its U.S. privacy notice says it may serve targeted ads to teenagers over 13. And a Character you create and share publicly can stay live after you delete your account.

Below is what the service collects, what survives deletion, what teens can actually do on the platform now, and where parental controls reach further than Character.AI's own settings.

Is Character.AI safe?

Character.AI is safer for teens than it was a year ago, but it's no more private than any other AI chatbot. Character.AI's U.S. State Privacy Notice lists which categories of personal information went to advertising providers over the past 12 months. Chat content isn't among them. Neither is voice, video, or sensitive information.

Your chats do reach vendors and service providers, including content moderation and analytics companies, and Character.AI's own affiliates.

Creating a public Character or a public Voice makes that content visible to other users by design. A private chat stays between you and the company. A public Character does not.

What happens to your Characters if you delete your account?

Some of it stays. The retention section of the privacy policy reserves the right to keep a Character you shared with other users, "even if you otherwise delete your data and your account." The stated reason is to avoid disrupting other people who use it.

Character.AI even advises not to put anything into the Character creation flow that you wouldn't want to stay there.

Deleting your account still removes your chats and your profile, but it doesn't necessarily remove the Character you built and shared.

Does Character.AI sell your data or show you ads?

Character.AI added more ad placements for free users by April 2026, and its CEO tied the decision to the cost of running the platform without outside investors.^[2] In December 2025, it also launched Charms, an in-app currency you earn through daily check-ins and quests, then spend on extras like additional image generations and ad-skip passes.^[3]

On selling, the U.S. State Privacy Notice says Character.AI disclosed five categories of personal information to advertising providers over the past 12 months, in transactions several state laws define as sales. The five are:

1. Identifiers, including your name, email address, phone number, username, IP address, and device identifiers
2. Demographic information, including your date of birth and age
3. Geolocation data, meaning approximate location based on your IP address
4. Device and network activity, including information about your device and how you interact with the service and its ads
5. Inferences drawn from your other personal information, such as your preferences and interests

Character.AI says that it doesn't sell personal information for money and that "Character.AI is not a data broker." Under California and similar laws, sharing data for targeted advertising counts as a sale whether or not money changes hands. And chat content, voice recordings, and sensitive information aren't in the sold column.

However, it clearly states using targeted ads for teens: "In some jurisdictions, we may provide targeted advertising to minors over the age of 13, which may include disclosures to third parties that are considered 'sales' in some jurisdictions." Character.AI adds that it obtains consent first where local law requires it.

That's a different posture from its competitors. OpenAI says it shows no ads at all to accounts it identifies as under 18. DeepSeek states it doesn't do targeted advertising. Learn more about whether ChatGPT is safe and how DeepSeek compares.

Can kids use Character.AI?

Officially, at 13 in the United States and 16 in the European Economic Area and the United Kingdom.

What teens can do changed completely in late 2025. On October 29, Character.AI announced it would remove open-ended chat for everyone under 18, and the rollout began November 24.^[4] In the weeks before, it cut teen chat time to two hours a day, then one. The company said it acted after regulators raised questions about what teens encounter in AI chat and how open-ended chat affects teens even when content controls work.

The legal pressure was substantial. In January 2026, Character.AI and Google settled five lawsuits brought by families who said their children were harmed after using AI chatbots, including cases involving teen suicide. The terms are confidential.^[5]

Five days after the chat cutoff, Character.AI launched Stories, a guided format built specifically for users under 18.^[6] You pick two or three Characters, choose a genre, write or generate a premise, then move the plot forward by picking from a set of options at each turn. Stories are replayable, and you can share them to the community Feed.

More followed. c.ai books turned classic literature into playable scenes in April 2026. c.ai series launched studio-made short vertical dramas in August. c.ai Comics arrived in September. Character.AI now describes itself as a connected entertainment platform, not a chatbot.

So the trade for teens isn't less to do. It's content on rails. Every new c.ai format above moves through choices the platform defines, instead of a conversation the teen steers.

As of September 2026, the company says its safety work includes self-harm detection that reads signals across a whole conversation rather than a single message, moderation notices that tell creators why content was actioned and let them appeal, and blocking that hides another user along with their Characters, Posts, Voices, and Scenes in both directions.^[7] It runs an in-house age-estimation model and partners with ConnectSafely, the Internet Watch Foundation, and StopNCII.

Regulators aren't finished. The EU is weighing rules that would cover AI companions. The House passed the KIDS Act in June 2026, which targets AI chatbots among other features. Meta's roughly $18 billion child-safety settlement, by contrast, excluded AI chatbots entirely.

Is giving your ID to Character.AI safe?

Most users are never asked for one. Character.AI runs age assurance on every account using signals like login information and platform activity.^[8] If it indicates 18 or older, nothing happens. If it points to under 18 and that's wrong, you verify with a selfie handled by Persona, a third-party vendor. An ID is requested only when the selfie is inconclusive, and Character.AI says it never sees the ID itself.

Character.AI says Persona keeps the IDs it collects for one week. Worth knowing that Persona's retention window is set by the company hiring it, and its own policy allows up to three years, so one week is Character.AI's choice rather than Persona's default.

We've covered the full flow in our breakdown of Character.AI age verification.

What parents can do

Character.AI's tool is called Parental Insights. It sends a weekly email covering how much time your teen spends on the platform, which Characters they interact with most, how much time they spend on each, and their subscription status. It excludes chat content entirely.^[9]

The only enrollment flow Character.AI documents starts with the teen. They open the parental insights tab in Preferences, enter your email, and send you the invitation. They can also request that you remove your email, but you have to approve the request before it takes effect.

So a parent can't turn it on themselves. Character.AI said in September 2026 that it rebuilt the tool with k-ID, without saying whether that changed who starts the process.

Either way, Character.AI's parental tool is limited. It reports time and Characters, not content, and it covers one app.

Device-level parental controls do more. They work whether or not your teen enrolls you, cover every app on the phone instead of one, and don't depend on a platform keeping its own tool current. Learn more about protecting your kids online.

How to protect your data on Character.AI

Here are some privacy settings you can change to protect your data on c.ai.

1. Turn off model training. 

• In the app: Settings > Data & Privacy, then deselect Improve the Model for Everyone. 
• On the web: Settings > Account > Manage Account & Data > Model Improvement, and deselect the same option. After that, new content won't be used to train Character.AI's generative models, though the company says it may still use your data to improve search, recommendations, and safety classifiers.

Character.AI's documents don't make it clear whether U.S. users have the same opt-out right. The Regional Privacy Disclosures say you can opt out of AI training at any time, with no geographic limit. The help article they point you to is addressed only to users in the European Economic Area and the UK. If you're in the U.S., try the settings path above and confirm the toggle is there.

2. Opt out of targeted advertising.

You can opt out of two things: targeted advertising, and the disclosures state laws treat as "sales" of your personal information. That's narrower than it sounds. Your data still goes to vendors, service providers, and affiliates, and Character.AI says the option is only available to residents of certain U.S. states, where local law requires it.

• On the web, use Your Privacy Choices in the footer, or go to Profile > Policies > Your Privacy Choices.
• In the app, it's Profile > Settings > Data and Privacy > Your Privacy Choices.

The catch is that you have to opt out on every device and browser. You can opt out on your laptop, but your phone may still be opted in.

3. Review what Memory has saved. Open the Memory screen from the chat header. You can edit auto-captured Facts, disable them, or remove side characters you'd rather not have stored.

4. Keep personal details out of public Characters. Anything in the creation flow may stay on the platform after you delete your account.

5. Treat every chat as a record. The privacy policy says Character.AI may disclose information in response to a warrant, subpoena, or court order. If you wouldn't want it read back to you, don't type it.

Bottom line: Is Character.AI safe?

Character.AI did something almost no competitor did. It took open-ended chat away from every user under 18, knowing that chat was the product.

It also built a lot to replace it. Stories, Books, Comics, and series give teens more to do than the coverage of the chat removal suggested. What they don't give is a conversation the teen controls.

The data side didn't move in the same direction. The company collects your chats, uploads, voice recordings, and location. Its Facts feature saves personal details as you talk and carries them between conversations. It discloses identifiers, your date of birth, and inferences about your interests to advertising providers, and says it may serve targeted ads to teenagers over 13. And it reserves the right to keep a Character you shared publicly after you've deleted everything else.

For an adult who keeps personal details out of the chat box, that's a reasonable trade. For a teenager, parental controls cover the part Character.AI's own settings don't.

Character.AI proved it will give up a feature to protect teens. It hasn't shown the same willingness to give up the data.

FAQs