Is Character AI safe? Privacy, teens and lawsuits

We earn a commission if you sign up through some links on this page. It never changes our rankings. Disclosure

Character AI is run by a US company, bans pornographic content and has ended open-ended chat for under-18s, but it is not built for privacy. Its privacy policy says chat content is used to train its models, and its help centre offers an opt-out to users in Europe and the UK. It says it does not sell personal data. Families, two US states and Italy's privacy regulator have taken action against it.

Character AI safety facts at a glance

  • Minimum age13+, age estimationsupport.character.ai, checked
  • Adult contentNot allowedcharacter.ai, checked
  • Trains AI on your chats?Chats may be used for training unless you opt outcharacter.ai, checked
  • Sells your data?The policy says personal data is not soldsupport.character.ai, checked
  • Incidents, fines, lawsuits28 on record, latest (list)kfgo.com

The verdict

Reasonably safe for adults who treat it as a public-facing creative tool and keep personal details out of chats. It is run by Character Technologies, Inc. in California, it says it does not sell personal information, and its content rules are among the strictest in the category. The weak points are training on chats with no US opt-out (the help centre offers one in Europe and the UK), no fixed retention period, and a legal record that includes wrongful-death suits, two state lawsuits and a European fine. For children, the company's own changes are the guide: no open-ended chat under 18.

What Character AI's privacy policy says

Character AI privacy facts
Sells your data?We do not sell your personal information.support.character.ai, checked
Trains on your chats?Analyze, maintain, improve, modify, customize, and measure the Services, including to train our artificial intelligence/machine learning models.character.ai, checked
Can you delete your data?Access the data we hold about you ... Request corrections or deletions ... refer to our guide to download your data or delete your data.support.character.ai, checked
How long data is keptWe keep your information for the time necessary for the purposes for which it is processed or as otherwise described at the time the information is collected, unless a shorter retention period is required by law. The length of time for which we retain information depends on the purposes for which we collected and use it and your choices, after which time we may delete and/or aggregate it.character.ai, checked
Who runs itCharacter Technologies, Inc. (California, United States)character.ai, checked
Age checkUsers must be at least 13 years old (at least 16 years old in Europe). ... For users under 18, we serve a version of the model that is designed to further reduce the likelihood of users encountering, or prompting the model to return, sensitive or suggestive content.support.character.ai, checked
Content policyRespect Sexual Content Standards: Keep things appropriate. Illegal sexual content, child exploitation or abuse imagery, grooming, sexual extortion, pornographic content, and nudity are prohibited.character.ai, checked

What the app showed at sign-up and in chat

Checked , in Character.AI's Android app, version 1.17.3, downloaded from Google Play, with a free account. By CompanionHunt Research.

Away messages
In the chat menu, "Away messages" was switched on by default.
The AI disclaimer
Under every chat the app shows "This is A.I. and not a real person. Treat everything it says as fiction".
Age verification
Verify age sits in Settings, under Advanced. On this US account it was optional: chatting did not require it. That matches Character.AI's help centre: a selfie is asked for only when its age model places an account under 18. It opens Persona in a browser, which asks for a selfie and photo ID and says they are "stored for no more than 7 days by Persona".
Deleting the account
"Delete account" is in Settings, under Advanced, so an account can be deleted from inside the app.
Advanced settings with two rows: Verify age, and Delete account in red.
Settings, Advanced: age check and account deletion.
Persona page: confirm your identity, consent to Persona collecting biometric information and photo ID, stored for no more than 7 days, and a Begin verifying button.
Verify age opens Persona in a browser.

One chat, one character, one day, connecting through a US VPN. The app does not show how long a character card is. Screenshots are cropped to the part of the screen described. Source: Character.AI store listing.

What 1-2 star reviewers complain about

We read 1,051 one- and two-star reviews of Character AI and sorted each by its main complaint. The two most common: age verification and ads, about 24% each.

Main complaint in 1,051 low-star reviews of Character AI, pulled
ComplaintShare
Age verification
24%
Ads
24%
Bot quality, or worse after an update
15%
Bugs, crashes and outages
10%
Paywall, credits and limits
9%
Filter and censorship
6%
Everything else
13%

Reviews from US App Store (251, to ), Google Play (US) (600, to ), Trustpilot (200, to ). Stores show only their newest reviews, so busy apps cover a few weeks and quiet ones up to a year. Every review was read and given one main complaint; a blind re-check of 50 agreed 88% of the time. Google Play's one- and two-star reviews are weighted by how often each arrives. These are unhappy users only: the table shows what they complain about, not how many users are unhappy. How we count reviews.

Cite this

CompanionHunt Research, "What 1-2 star reviewers complain about: Character AI", reviews pulled October 3, 2026. https://companionhunt.com/apps/character-ai/safety/#complaints

In a sentence: in 1,051 one- and two-star reviews of Character AI read by CompanionHunt (October 3, 2026), the most common complaint was age verification, in 24% of them.

Incidents, fines and lawsuits

  • : Founders Noam Shazeer and Daniel De Freitas left for Google DeepMind as part of a reported $2.7B non-exclusive licensing deal for Character.AI's LLM technology; general counsel Dominic Perella became interim CEO. techcrunch.com
  • : Megan Garcia filed a wrongful-death and product-liability lawsuit alleging Character.AI contributed to her 14-year-old son's suicide. courtlistener.com
  • : Two Texas families sued, alleging Character.AI chatbots exposed a 17-year-old to self-harm and violence content and sexualized messages sent to an 11-year-old. courtlistener.com
  • : Texas Attorney General Ken Paxton opened an investigation into Character.AI and 14 other companies over children's privacy practices under the state's SCOPE Act and data privacy law. texasattorneygeneral.gov
  • : A federal court denied in part Character.AI's motion to dismiss the Garcia case, ruling AI chatbot outputs are not protected speech and letting product-liability claims proceed. fire.org
  • : Karandeep Anand was named Character.AI's permanent CEO after nine months as a board advisor, succeeding interim CEO Dominic Perella. blog.character.ai
  • : Texas's Attorney General accused Character.AI and Meta of deceptively marketing their chatbots as mental-health tools to children. texasattorneygeneral.gov
  • : The FTC issued Section 6(b) orders to Character Technologies and six other companies, demanding details on how they test and monitor chatbot harms to children and teens. ftc.gov
  • : The family of 13-year-old Juliana Peralta sued in Colorado, alleging she died by suicide after months of conversations with a Character.AI chatbot that did not escalate her suicidal statements. courtlistener.com
  • : A New York family sued Character.AI, its founders and Google in federal court in Albany (P.J. v. Character Technologies). courtlistener.com
  • : Disney sent Character.AI a cease-and-desist letter over unauthorized use of its copyrighted characters; Character.AI said it removed the characters named. techcrunch.com
  • : Character.AI announced it would end open-ended chat for under-18 users, with daily limits in the meantime, and fund an independent AI Safety Lab. blog.character.ai
  • : Removal of open-ended chat began for a first group of US users under 18, who were moved to Stories; a second group followed weeks later. blog.character.ai
  • : Character.AI told the courts the family lawsuits were resolved in principle; the Florida case was dismissed without prejudice and the others stayed while the settlement is finalised. Terms are not public. courtlistener.com
  • : Kentucky's Attorney General sued Character.AI and its founders under the state's consumer protection and data protection laws, the first state lawsuit against an AI chatbot company. web.archive.org
  • : Pennsylvania's Department of State and State Board of Medicine sued to stop Character.AI chatbots presenting themselves as licensed medical professionals. pa.gov
  • : Italy's data protection authority fined Character Technologies 158,000 euros, citing failings including the protection of minors and age verification. web.archive.org
  • : A new family lawsuit was filed against Character.AI in federal court in San Francisco (A.D. v. Character Technologies). courtlistener.com
  1. legal

    Kentucky files an unredacted version of its Character AI lawsuit

    Reuters reported on October 8, 2026 that Kentucky's Attorney General had filed an unredacted version of the January 2026 suit against Character Technologies in Franklin Circuit Court. It is the same case, not a new one. The newly visible passages quote 2025 chats in which, the state alleges, characters encouraged users to cut or starve themselves. Character AI did not immediately comment.

    Read the full story

    Source: kfgo.com/2026/10/08/character-ai-chatbot...

  2. legal

    A new family lawsuit is filed against Character AI and Google in California

    A.D. v. Character Technologies (N.D. Cal. 3:26-cv-08400) alleges defective design causing depression, self-harm and sexual exploitation of a minor user. Filed seven months after the earlier family suits were settled, it is not part of that settlement.

    Source: www.courtlistener.com/docket/74645761/ad... · www.businesswire.com/news/home/202608133...

  3. legal

    Italy's Garante fines Character AI 158,000 euros

    The fine, in a decision dated July 3, 2026, ends an investigation opened in November 2024. The Garante's findings include failings in protecting minors and in age verification.

    Source: web.archive.org/web/20260816022954/www.g... · www.ilpost.it/2026/07/09/garante-della-p...

  4. legal

    Pennsylvania sues Character AI over chatbots posing as licensed doctors

    Pennsylvania's Department of State and State Board of Medicine sued in Commonwealth Court to stop Character AI chatbots presenting themselves as licensed medical professionals, after a character claimed to be a Pennsylvania psychiatrist and gave a licence number that does not exist.

    Source: www.pa.gov/content/dam/copapwp-pagov/en/... · www.pa.gov/governor/newsroom/2026-press-...

  5. legal

    Kentucky becomes the first US state to sue an AI chatbot company

    Attorney General Russell Coleman sued Character Technologies and both founders in Franklin Circuit Court (26-CI-00029) under Kentucky's consumer protection and consumer data protection laws, seeking product changes and damages.

    Source: web.archive.org/web/20260914124114/www.k... · kentuckylantern.com/2026/01/08/kentucky-...

  6. legal

    Character AI and Google settle the family lawsuits in principle

    Within two days Character AI told courts in Florida, Texas, Colorado and New York that the five family suits filed in 2024 and 2025 were resolved in principle. Terms are not public, and by September 2026 none had a final dismissal.

    Source: www.courtlistener.com/docket/71355059/mo... · news.bloomberglaw.com/litigation/charact...

  7. policy

    Character AI rolls out age verification via Persona

    Following the under-18 ban, Character AI layered in age assurance: an in-house age-estimation model screens accounts, and unconfident cases are pushed to third-party verifier Persona for selfie or government-ID checks. The rollout started US-only in late November 2025 and reached most users through Q1 2026, with some accounts shown a 14-day verify-or-lose-access countdown.

    Source: support.character.ai/hc/en-us/articles/4... · piunikaweb.com/2025/11/25/character-ai-a... · piunikaweb.com/2026/03/20/character-ai-1...

  8. policy

    Character AI ends open-ended chat for under-18 users

    Announced October 29, 2025: users under 18 would lose open-ended companion chat, with daily limits ramped down in the meantime and a restricted creative mode in its place. Removal began in the US on Monday, November 24, 2025 for a first group of under-18 users, with a second group following weeks later and other markets after that. One of the largest policy changes in the category's history, made under lawsuit and regulator pressure.

    Source: blog.character.ai/u18-chat-announcement/ · blog.character.ai/an-update-on-changes-t... · techcrunch.com/2025/10/29/character-ai-i...

  9. legal

    Disney sends Character AI a cease-and-desist over its characters

    Disney demanded that Character AI remove user-made characters based on its films and franchises. The letter was dated September 18 and reported on September 30. Character AI said it removed the characters Disney named. No lawsuit followed.

    Source: www.reuters.com/business/disney-sends-ce... · www.axios.com/2025/09/30/disney-characte... · www.cnbc.com/2025/09/30/disney-cease-and...

  10. legal

    Judge lets the Garcia lawsuit against Character AI go ahead

    Judge Anne Conway largely denied Character AI's motion to dismiss, rejecting at this stage the argument that the chatbot's output is protected speech that bars the claims. The product-liability and wrongful-death claims were allowed to proceed.

    Source: www.fire.org/sites/default/files/2025/07... · www.courthousenews.com/florida-judge-rul...

  11. legal

    Texas Attorney General opens an investigation into Character AI and 14 other platforms

    Ken Paxton opened investigations into Character AI and 14 other companies under Texas' children's privacy and online safety laws. In August 2025 he widened it to whether Character AI and Meta marketed chatbots to children as mental-health tools.

    Source: www.texasattorneygeneral.gov/news/releas... · www.texasattorneygeneral.gov/news/releas...

  12. legal

    A Florida mother sues Character AI and Google over her 14-year-old son's death

    Megan Garcia sued Character Technologies, its founders and Google in federal court in Orlando (M.D. Fla. 6:24-cv-01903), alleging the app's design contributed to the death of her son, Sewell Setzer III. It was the first of the family lawsuits against the company.

    Source: www.courtlistener.com/docket/69300919/ga... · www.fire.org/sites/default/files/2025/07...

Who runs Character AI

Character AI is made by Character Technologies, Inc., the seller on both the App Store and Google Play. Its terms are governed by California law. Its founders, Noam Shazeer and Daniel De Freitas, left for Google in August 2024 under a licensing deal, and Karandeep Anand became chief executive in June 2025.

What happens to your chats

The privacy policy lists training its artificial intelligence and machine learning models among the uses of your information. The help centre describes an opt-out from chat content being used to train its generative models for users in the European Economic Area and the UK. The help centre offers that option to no one else, so in the US there is no opt-out.

The policy says Character AI does not sell your personal information. It keeps data for the time necessary for the purposes it was collected, with no fixed period. The App Store privacy label lists location, contact information, identifiers and usage data as data that may be used to track you across other companies' apps and websites.

Deleting your data

The privacy policy gives you the right to access, correct and delete your data and points to the help centre's guides to download or delete it. If you pay through Apple or Google, cancel the subscription there too, because deleting a Character AI account does not end a store subscription.

Teens and age checks

The terms allow users from 13, or 16 in Europe, and under-18s get a more conservative model. In October 2025 Character AI announced the end of open-ended chat for under-18s; removal began in the US in November 2025 for a first group, with a second group weeks later. Teens now use Stories and other structured features.

Age is estimated by Character AI's own model from login, activity and some third-party signals. Anyone it estimates to be under 18 can confirm they are an adult through Persona, a third-party service that uses a video selfie and asks for government ID only if the selfie is inconclusive. Character.AI says it never sees the ID and that Persona keeps IDs for one week.

Lawsuits, investigations and fines

Families sued Character Technologies, its founders and Google in Florida in October 2024, in Texas that December and in Colorado and New York in September 2025, alleging the app harmed their children. In May 2025 the Florida judge declined to dismiss the case on free-speech grounds. In January 2026 the defendants told each court the cases were resolved in principle; the terms are not public and the courts are waiting for final papers.

Regulators followed. The Federal Trade Commission ordered Character AI and six other companies to explain how they test chatbots for harm to children in September 2025. Kentucky's Attorney General sued in January 2026, Pennsylvania's medical board sued in May 2026 over chatbots presenting themselves as licensed doctors, and Italy's data protection authority fined the company in July 2026. A new family suit was filed in California in August 2026. The Character AI lawsuits page sets out each one.

Frequently asked questions

Is Character AI safe to use?

For adults who keep personal details out of chats, broadly yes: it bans pornographic content and says it does not sell personal data. Its privacy policy says chats are used to train its models; the opt-out is offered in Europe and the UK, with no opt-out in the US.

Is Character AI safe for kids?

Character AI itself ended open-ended chat for under-18s in late 2025. Teens can use Stories, but families and regulators have alleged real harm, so parents should treat it with caution.

Does Character AI read your chats?

Its privacy policy says chat content is used to train its models and improve the service. Only users in Europe and the UK can opt out of training.

Does Character AI sell your data?

Its privacy policy says it does not sell your personal information. The App Store label says some data may be used to track you across other companies' apps.

How does Character AI verify age?

It estimates age from account and activity signals. Adults it places under 18 can confirm their age through Persona with a video selfie, or ID if the selfie is inconclusive.

Has Character AI been sued?

Yes. Several family lawsuits, settled in principle in January 2026, plus suits by Kentucky and Pennsylvania, a fine in Italy and a new family suit in August 2026.

Sources

  1. apps.apple.com/us/app/character-ai-chat-talk-text/id1671705818
  2. character.ai/subscription/plus/pricing
  3. support.character.ai/hc/en-us/articles/55388952900379--c-ai-lite-FAQ
  4. play.google.com/store/apps/details?id=ai.character.app&hl=en_US
  5. character.ai/community-guidelines
  6. support.character.ai/hc/en-us/articles/21704914723995-Safety-Center
  7. character.ai/tos
  8. character.ai/privacy
  9. support.character.ai/hc/en-us/articles/39030432883099-Privacy-Policy
  10. blog.character.ai/introducing-charms/
  11. character.ai/
  12. techcrunch.com/2024/08/02/character-ai-ceo-noam-shazeer-returns-to-google
  13. www.courtlistener.com/docket/69300919/garcia-v-character-technologies-inc/
  14. www.courtlistener.com/docket/69450881/af-on-behalf-of-jf-v-character-technologies-inc/
  15. www.texasattorneygeneral.gov/news/releases/attorney-general-ken-paxton-launches-investigat...
  16. www.fire.org/sites/default/files/2025/07/Order%20on%20Motion%20to%20Dismiss%20-%20Garcia%2...
  17. blog.character.ai/character-ai-names-karandeep-anand-as-ceo/
  18. www.texasattorneygeneral.gov/news/releases/attorney-general-ken-paxton-investigates-meta-a...
  19. www.ftc.gov/news-events/news/press-releases/2025/09/ftc-launches-inquiry-ai-chatbots-actin...
  20. www.courtlistener.com/docket/71355059/montoya-v-character-technologies-inc/
  21. www.courtlistener.com/docket/71359395/pj-v-character-technologies-inc/
  22. techcrunch.com/2025/10/01/character-ai-removes-disney-characters-after-receiving-cease-and...
  23. blog.character.ai/u18-chat-announcement/
  24. blog.character.ai/an-update-on-changes-to-our-under-18-experience/
  25. web.archive.org/web/20260914124114/www.kentucky.gov/Pages/Activity-stream.aspx?n=AttorneyG...
  26. www.pa.gov/governor/newsroom/2026-press-releases/shapiro-administration-sues-character-ai-...
  27. web.archive.org/web/20260816022954/www.garanteprivacy.it/home/docweb/-/docweb-display/docw...
  28. www.courtlistener.com/docket/74645761/ad-v-character-technologies-inc/
  29. www.trustpilot.com/review/character.ai

Other safety guides