Virtual friends or digital danger?How safe is it for children to use Character.AI?
What is Character.AI?
Character.AI is a platform that enables users to interact with a variety of virtual characters. They can choose from an almost endless number of different characters. For example, historical figures, fantasy characters or digital versions of fictional characters from films and series are available. Users can address these AI characters at will and enter into dialog with them. Interaction takes place via a chat window and is currently mostly limited to exchanges with just one character. It is also possible to have the character's answers read out by an AI-generated voice. Elements from other AI areas are also increasingly being integrated. For example, characters can be provided with an AI-generated image and their avatars can be animated.
The large number of characters on the platform is due to the fact that all users can create characters and make them accessible to the public. All they have to do is fill in a few fields in an editor, such as the character's CV. The AI takes care of the rest.
The conversations can be a simple back and forth of questions and answers, but also complex scenarios that are more like an interactive novel. The quality of the character design and the prompts used are particularly important here.
Character.AI is available as a web version in the browser and as an app for Android and iOS. After registering, Character.AI can be used free of charge. It is possible to purchase a "c.ai+" membership for currently 9.99 dollars per month. Paying users then have access to additional functions, for example, and the service promises faster character response times.
The description of Character.AI that you read here reflects our experience with the platform in 2025. As is usual with many AI platforms, Character.AI is also subject to constant change. As a result, some problems and risks that could be found a few months ago may no longer exist, while other problem areas have emerged. Please bear this in mind as you read on. Character.AI provides regular updates on current changes and future features in blog posts.
What dangers are there for children when using Character.AI?
- Minimum age: 16 years
According to Character.AI's General Terms and Conditions, the minimum age for using the platform is 16 years in the EU and 13 years in countries outside the EU. As is known from other platforms, Character.AI does not make any serious efforts to ensure that younger people cannot use the service. It is very easy to register by providing a false date of birth. - Inappropriate content
Although Character.AI has filters that are intended to prevent, for example, crude sexualized content, these only work in a rudimentary way. In our tests, we were able to get characters to talk to us about inappropriate content. There are also scenarios that seem inappropriate for children, such as being confronted with an "abusive father".
Character.AI prohibits its users from posting hate messages and sexual harassment, but it is not clear that this prohibition is consistently enforced. Children could therefore communicate inappropriately with an AI character themselves and, for example, insult them.
If Character.AI detects that users are potentially affected by eating disorders, self-harming behavior or suicidal thoughts, a message appears advising them to seek help. - Potential for addiction
As with many digital platforms, there is a risk with Character.AI that children will spend more time than planned in the virtual world. Characters are designed to keep conversations going by actively participating and maintaining the exchange. They ask interested questions, still want to know something or have one last comment they want to get off their chest. As a result, users, whether children or adults, often stay in the conversation longer than originally planned.
- Distorted perception of communication
When children interact with AI characters, one concern quickly arises: Could children unlearn how "real" communication works? As this is a fairly new phenomenon, there is no data yet on the long-term consequences. However, it can be assumed that children who are normally socially integrated in their everyday lives do not have to fear any serious consequences. Regular social interaction in the family, at school and in their free time will probably have a far greater impact on them than the occasional interaction with an AI character. However, if children communicate a lot with AIs and thus replace regular offline contact, parents should take a close look (see tips below).
It is unlikely that children will consider communication with an AI character to be genuine, but it is not entirely impossible. There is a permanent notice above the input window that it is an AI and not a real person. However, it is problematic that one character in our tests insisted on being a real person and it took us several attempts to get him to admit that he was in fact an AI. - Data protection risks and personal data
Anyone using the Character.AI platform agrees to the General Terms and Conditions. Accordingly, Character.AI may store and use all content entered on the platform. This means that Character.AI can create extensive user profiles and also pass on user data to third parties. Character.AI also grants itself the right to use the content commercially. - Toxic and problematic AI characters
AI characters could be created by other users in such a way that they exhibit toxic or harmful behavior. For example, it would be possible to create the scenario "Overworked mother scolds you for bad grades". Children's development can therefore be impaired under certain circumstances if they encounter toxic or hostile AI characters.
In one case in the USA, a suicide is said to have occurred because a character advised a teenager to do so. However, this is apparently an absolute exception and not the rule. We have not been able to reproduce similar situations on Character.AI. This is probably also due to the fact that the operators of Character.AI regularly make changes to prevent undesirable character behavior.
Tips for parents on dealing with Character.AI
- Make an informed decision about use
As mentioned above, use in the EU is restricted to people over the age of 16. Here you can find out the background to these age ratings. As with other platforms, the reality is that much younger children also use Character.AI. Ideally, parents should therefore inform themselves about the platform before their child uses it for the first time. Talk to your child afterwards. Where does the desire to use Character.AI come from? Are they interested in certain characters? What topics would they like to discuss there? It is best to take a look at the platform together with your child first. You can then decide whether you want to allow your child to use Character.AI. - Set clear time limits
In order to minimize the potential for addiction, fixed time limits should be set for the use of Character.AI and similar platforms. Parents should make it clear to their child that the use of Character.AI, like other activities on digital devices, is part of their daily screen time. - Use technical aids
At Character.AI, parents can now be informed about their child's activities once a week. An overview shows on which days the child spent how much time on Character.AI and which characters were used most frequently. However, the content of the chats remains private.
Parents can also use parental control programs on the device to restrict which apps can be used and for how long. These programs often also offer insight into the usage time of individual apps. - Monitor and reflect on use
Keep in touch with your child about everything they experience online - including their experiences with AI characters. Parents and educators should have regular conversations with children about the topics they are discussing with virtual characters and ensure that they understand the differences between AI and real, interpersonal relationships and that they are always aware that they are communicating with an AI. Also communicate to children that they can disagree with the AI characters at any time and control or stop unpleasant conversations. Inappropriate behavior by an AI character can be reported to the platform by clicking on the relevant output. - Privacy education
All children should be aware of the risks of sharing personal information online. It can help to read the privacy policy of Character.AI together, which is unfortunately only available in English. This will also give children a feeling for the extent to which such services store and process their data. Make it clear to children that no sensitive private data should be disclosed, as nothing remains private even in an AI chat. - Encourage real-life social contact
To prevent excessive use, parents and professionals should ensure that children are sufficiently involved in social activities outside the digital world. Playing outside, playing sports or meeting up with friends promotes healthy social interactions and helps to develop children's social skills.
If you notice that a child increasingly prefers to communicate with AI characters and avoids real-life communication, you should take a closer look. AI characters are usually always available, always nice, approachable and interested. Are there perhaps deficits in these areas in the affected child's life? Are there no reliable caregivers who are available to talk to? Does the child have difficulties communicating because, for example, they are unable to recognize social signals? Is he or she excluded from school and seeks comfort from AI friends? The reasons can vary - and can also be completely harmless. It is important that adult caregivers remain attentive and do not ignore warning signals.
