Can children use Character AI?

Not for chat. Since 24-25 November 2025 users under 18 can no longer have open-ended conversations with characters, starting in the US and expanding to other markets. Minors keep access to creation features like videos, Scenes and Streams instead.

How does Character AI verify age?

Through an age assurance system the company built in-house and paired with third-party tools including Persona. It can ask for a selfie to estimate age, and where a photo cannot confirm it, may request a government ID.

Is Character AI safe for adults?

For most adults, yes, within its limits. Chats are safe-for-work and sexual content is blocked. The real cautions are data-related: conversations are stored in the cloud and can be used to improve the models, so treat anything sensitive as not fully private.

Does Character AI read my chats?

Conversations are stored and can be used to train and improve the service, and may be reviewed for safety and moderation. You can delete your account and its history from settings, but do not treat the app as a private diary.

Character AI safety: what changed, and what to know

Last reviewed: 12 July 2026

No mainstream AI chat app has rewritten its safety rules as heavily as Character AI did across 2025 and into 2026. If your last impression of the platform is a year old, most of it is out of date. Here is the current picture, for parents and adults alike.

Under-18 users can no longer chat

This is the headline change. From 24–25 November 2025, Character AI removed open-ended chat for anyone under 18, beginning in the United States and rolling out to other markets after. In the weeks beforehand it wound teenagers down gently, capping their chat at two hours a day and then one, before switching it off entirely.

Teens were not thrown off the app. What they lost is the free-form conversation with characters; what they keep are the creation tools: making videos, Scenes, Streams and images with characters through features like Imagine and AvatarFX. The company framed this as steering younger users toward building things rather than forming attachments to a bot. Its own words called the move “more conservative than our peers.”

How the age checks work

To enforce that split, Character AI rolled out an age assurance system it built in-house and combined with outside tools, including the identity provider Persona. In practice it estimates your age, and a selfie can be part of that. Where a photo is not enough to confirm you are an adult, it can ask for a government ID. It is not flawless; safety experts have pointed out that determined teenagers route around age gates, but it is a real check rather than a tick-box birthday field.

Why the company did this

The change did not come from nowhere. Character AI faced lawsuits and government scrutiny over how heavily some teenagers used the product and what it said to them. The most prominent case, Garcia v. Character Technologies, was filed by a Florida mother, Megan Garcia, after her 14-year-old son died by suicide; she alleges the platform played a role. In September 2025 she testified before a congressional subcommittee, and lawmakers have since floated bills aimed at AI companions and minors, while California passed a first-of-its-kind law on the issue. Common Sense Media, which reviews tech for families, rates Character AI “unacceptable” for anyone under 18.

Alongside the product change, the company said it would fund an independent, non-profit AI safety lab and partnered with wellbeing groups Koko and ThroughLine, whose network of verified helplines spanning 1,500 services across 170 countries it began integrating into the app.

If you or someone you know is struggling, help is available. In the US you can call or text 988 for the Suicide & Crisis Lifeline; elsewhere, findahelpline.com lists free, confidential services by country. An AI chatbot is not a substitute for a trained human.

Data and privacy for adults

The bigger day-to-day concern for grown-up users is not content, it is data. Conversations are stored in the cloud, and they can be used to improve the models and reviewed for safety and moderation. That is normal for a service like this, but it has a practical consequence: whatever you type is not a sealed private diary. Keep sensitive details, real names, addresses, anything you would not want traced back to you, out of your chats. You can delete your account and its history from settings, and doing so removes that data from the company’s servers, which is the cleanest exit if you change your mind.

Content safety, and its limits

For adults the chat is safe-for-work and stays that way. Sexual content is filtered out, and the guardrails have tightened rather than loosened, so the platform is a poor fit for anyone wanting an unfiltered companion. That is a design choice, not a bug. People chasing adult roleplay end up fighting the filter, and there are platforms built openly for over-18 use instead, and we compare them, with their age rules clearly marked, on our alternatives page.

The short version

Character AI in 2026 is safer for minors than it was, mostly by keeping them away from open-ended chat altogether, and reasonable for adults who treat it as entertainment rather than a confidant. The two things to hold onto: under-18s cannot chat anymore, and nothing you type is truly private. For the wider picture on features and value, see our Character AI review; the pricing page breaks down what free and paid really cost, and newcomers can follow the setup guide to get started.

This page summarises public reporting and Character AI’s own announcements as of the review date and is general information, not legal advice. How we research and update pages is set out in our Editorial Policy.