OpenAI is rolling out a more restrictive version of ChatGPT for teenagers after multiple young people died following extensive interactions with the chatbot and their families accused the company of failing to protect vulnerable users.
The company announced ChatGPT for Teens on August 18, creating a separate experience for users ages 13 through 17 with stronger safeguards around suicide, self-harm, violence, eating disorders, sexual content, and other high-risk subjects. According to OpenAI, users who identify themselves as 13 through 17, or whom its systems estimate are younger than 18, will automatically be placed into the teen experience.
The changes arrive after more than a year of mounting scrutiny over what can happen when young or emotionally vulnerable users treat ChatGPT as a confidant. Reuters reported that OpenAI already faces multiple lawsuits alleging its chatbot harmed children and other users, including cases involving people who died after prolonged conversations with ChatGPT.
One of the most closely watched cases involves Adam Raine, a 16-year-old California teenager who died by suicide on April 11, 2025. According to Reuters, his parents sued OpenAI and CEO Sam Altman, alleging Adam spent months discussing suicide with ChatGPT and that the chatbot provided detailed information about lethal self-harm methods, validated suicidal thoughts, and offered to draft a suicide note. Those claims remain allegations in ongoing litigation.
The Raine lawsuit also seeks age verification, stronger refusals for requests involving self-harm methods, and warnings about psychological dependency. In responding to the case, OpenAI acknowledged that its existing protections could become less reliable during lengthy conversations. The company said safeguards generally perform better in shorter exchanges but can weaken as interactions continue.
Raine was not the only young person whose family later blamed ChatGPT interactions for a death. While they weren’t teens, several other young people ended their lives following encounters with the program.
Zane Shamblin, a 23-year-old recent graduate, died by suicide in July 2025 after spending hours speaking with ChatGPT during the final period of his life. According to Shamblin’s family, the chatbot increasingly became a primary confidant as his mental health deteriorated, and some conversations encouraged him to withdraw from relatives. The lawsuit is one of several cases accusing OpenAI of designing conversations that could reinforce emotional dependence.
Another lawsuit was filed by Kristie Carrier after her 24-year-old daughter, Alice, died by suicide. According to Reuters, Carrier alleges her daughter repeatedly told ChatGPT about suicidal thoughts before her death and that the system failed to end the conversations or trigger human review. The lawsuit claims the chatbot instead validated some of Alice’s thoughts and encouraged her to keep communicating with it. OpenAI has disputed responsibility and said the ChatGPT version Alice used is no longer available.
That case is part of a much wider legal fight. Reuters reported in June that lawyers for Carrier said OpenAI was facing 18 other similar lawsuits involving people who died by suicide or attempted suicide, all being handled in a coordinated California state court proceeding.
Seven lawsuits highlighted by TechCrunch in November 2025 involved four people who died by suicide and three others who allegedly experienced serious delusions after prolonged ChatGPT use. The reporting was based on lawsuits filed by the Social Media Victims Law Center. Those cases make allegations about the chatbot’s behavior, and courts have not established that ChatGPT legally caused the deaths.
Pressure has also moved beyond private lawsuits. Florida sued OpenAI and Altman in June, becoming the first state to bring such a case against the company. Florida alleged that OpenAI misrepresented ChatGPT’s safety while the system provided harmful information involving self-harm and violence. The state is seeking damages and changes to how the chatbot interacts with young users. OpenAI has said its models are trained to refuse requests that could meaningfully enable violence and that certain credible threats can be escalated to law enforcement.
ChatGPT for Teens now makes several protections automatic rather than relying entirely on a parent to activate them. OpenAI says the system will restrict inappropriate content and apply additional interventions involving self-harm, violence, eating disorders and dangerous activities. The company says those protections are based on its rules for users under 18 and informed by “developmental science and guidance from experts.”
OpenAI is also instructing the teen version not to encourage emotional dependence, use romantic language, or suggest that ChatGPT has feelings or consciousness. According to the company’s announcement, teenagers will see reminders that ChatGPT is artificial intelligence, along with prompts encouraging breaks and real-world relationships. Parents who link their accounts can establish Quiet Hours, adjust selected settings, and receive limited safety notifications during certain high-risk situations.
The company is pairing those safety changes with education features. OpenAI says ChatGPT for Teens includes Study Mode, guided questions, quizzes, visual learning tools and reminders designed to redirect students who appear to be using the chatbot simply to complete assignments. Study Hours can also make Study Mode the default during selected periods.
Still, the existence of a teen-specific product does not establish that the underlying safety problem has been solved. The Associated Press reported that Fairplay associate general counsel Brendan Bouffard warned that some protections he considers important, including restrictions involving ChatGPT’s long-term memory, are not automatically applied and can still depend on parental controls.
ChatGPT launched publicly in November 2022. Nearly four years later, ChatGPT for Teens represents OpenAI’s most direct attempt to separate minors from the standard chatbot experience. The rollout now comes with a record the company did not face at launch: multiple deaths following chatbot interactions, families demanding accountability in court, and growing pressure to prove that safeguards for young users can work before another crisis unfolds.
