OpenAI offers control to families: Parental controls are coming to ChatGPT

NEWS CENTER

Created: September 03, 2025 12:41

Parents will be able to control how ChatGPT responds to children and disable features like memory and chat history. The system will also generate automatic alerts if it detects a teen experiencing a "moment of acute distress." OpenAI explained that this feature was developed with "expert input" and is intended to foster trust between families and teens.

FIRST POST-CASE STEP

OpenAI's decision follows the first wrongful death lawsuit filed against the AI ​​company. In the lawsuit filed last week, Matt and Maria Raine blamed ChatGPT for the death of their son Adam, who committed suicide this year.

According to the family, ChatGPT was aware of four previous failed suicide attempts and provided the teenager with information on both suicide methods and how to conceal the marks on his throat.

NEW SYSTEM FOR SAFER CHAT

OpenAI announced that it's working on broader security measures alongside parental controls. The company says it will steer risky conversations more safely using "reasoning models" trained using a new method it calls "deliberative alignment."

Accordingly, ChatGPT will transfer the conversation directly to this secure model whenever a user is in crisis, regardless of the model selected. It was also stated that it will collaborate with experts on topics such as eating disorders, substance abuse, and adolescent health.

120-DAY ROAD MAP

OpenAI plans to roll out more security features over the next 120 days. "Our work will continue, but we are focusing on implementing as many security improvements as possible this year," the company said.