Is Character AI Safe? Privacy, Data and Age Policy Explained

Last updated: 15 July 2026

"Is it safe" is a fair question to ask before you hand an app your conversations, and with Character AI the honest answer has layers: who owns it, what happened that forced its safety policy to change, and what that means for your own data today. Here's the full picture, sourced rather than assumed.

Who actually runs it

Character AI was founded in November 2021 by Noam Shazeer and Daniel De Freitas, two engineers who left Google after building the technology behind what became Bard. In August 2024, Google paid roughly $2.7 billion for a non-exclusive licence to the company's AI models, and both founders returned to Google's DeepMind unit along with around 30 of their researchers. Character AI kept operating as an independent company, now under CEO Karandeep Anand, but its original technical leadership is gone, and its core technology is licensed to Google, one of the largest AI and advertising companies in the world. That's worth knowing before you decide how private a conversation on the platform really feels.

The safety history that shaped today's rules

Character AI's current age policy didn't appear out of caution alone. In October 2024, Megan Garcia filed a wrongful-death lawsuit after her 14-year-old son, Sewell Setzer III, died by suicide following months of an intense relationship with a Character AI chatbot. A federal judge in Florida allowed the case to proceed in May 2025, a first-of-its-kind ruling that a conversational AI could plausibly owe a duty of care to a minor user. Several related family lawsuits followed in Colorado, New York and Texas. In January 2026, Character AI, its founders and Google agreed to settle Garcia's case along with four related suits, on confidential terms and without admitting liability, as CBS News reported at the time. Litigation from other families, and separate state attorney-general inquiries, continues. If this subject affects you personally, the 988 Suicide & Crisis Lifeline is available by call or text in the US, and Samaritans can be reached free on 116 123 in the UK.

The product response came before the settlement. In late October 2025, Character AI announced it would remove open-ended, freeform chat entirely for any account identified as under 18, effective November 25, 2025, backed by an age-assurance system built with the verification provider Persona. Most adults are cleared silently from account signals alone; a selfie is only requested if the system suspects a minor account, and an ID only as an absolute last resort, which Persona states it deletes shortly after processing.

What that means for your own data

Two things are worth separating. The age-assurance system itself is relatively conservative with data: most users never see it, and biometric data, when it is collected at all, sits with Persona rather than Character AI. Separately, general platform data — your chats, your account activity, your usage patterns — trains and improves a product whose underlying model is now licensed to Google, which is a different privacy question from age verification and worth reading the actual privacy policy over rather than assuming either way.

A practical checklist

Run this on Character AI, or on whatever you're considering instead.

• Read the actual privacy policy, and search it for "share," "third parties," and "training."
• Check whether a working account-deletion option exists before you get attached to a character.
• If age assurance matters to you, know that it's triggered by suspicion signals, not a one-time selfie at signup.
• Understand that no verified-adult mode enables explicit content here — the content ceiling is the same for every account, which our breakdown of the NSFW policy covers in full.
• Don't share anything you couldn't live with surfacing in a data breach or legal discovery process.

The verdict

On safety in the child-protection sense, Character AI has moved, under real legal and public pressure, toward a more responsible position than it held in 2023 and 2024. On privacy in the everyday sense, it sits where most large consumer AI products sit: reasonably documented, licensed to a tech giant, and worth reading closely rather than trusting on the strength of its reputation. If you're specifically weighing it against an 18+ platform built around adult users from the outset, our OurDream AI safety breakdown walks through the same questions for this site, and our head-to-head comparison puts both side by side.

Frequently asked questions