OpenAI introduces parental control for ChatGPT

Children and young people can no longer engage in sexual, romantic, or violent role-playing games with ChatGPT. This is just one part of the new safety measures for minors that OpenAI is introducing. They apply with immediate effect and also in Germany. OpenAI has also provided a website for parents with help and ideas for dealing with AI for young people.

In principle, a young person's ChatGPT account must first be linked to that of an adult family member. Only then do the youth protection regulations apply. In addition to role-playing games, other inappropriate content is also filtered and no longer displayed in the chatbot. This includes viral challenges and extreme beauty ideals, writes OpenAI in the blog post.

Parents can also control the settings of young people's accounts. For example, it is possible to switch off the memory function, which remembers previous conversations and other information about a user and adapts content based on this. Voice mode can also be switched on or off. This also applies to image generation and the forwarding of data to OpenAI.

Parents can set blocking times so that children and young people are not constantly chatting at night or at school.

OpenAI protects privacy—in its own way

“OpenAI is aware that young people also use ChatGPT in difficult times and has therefore set up a system for notifying parents if serious problems are suspected.” The background to this are allegations that young people are having overly close alleged relationships with the chatbot and a lawsuit in which parents accuse OpenAI of assisted suicide. According to OpenAI, it worked with experts from the fields of mental health, education, and counseling on the messaging function. However, parents do not get the chance to read all the chats their children have. The privacy of young people is taken seriously, writes OpenAI. It is not a standard setting that the chats of young people are not sent to OpenAI. This form of privacy is therefore not meant.

In the future, OpenAI is working on automatically recognizing users under the age of 18. Parental control measures will then apply to them.

ChatGPT product manager Nick Turley recently wrote on X that new security measures will be made available for all ChatGPT users. The model changes in the background for particularly sensitive and emotional topics. There is now a GPT-5 specially designed for such cases, which is labeled chat-safety and says so when asked.

(emw)