OpenAI to Introduce Parental Controls for ChatGPT After Teen’s Suicide Linked to AI Conversations

OpenAI has announced it is developing new parental controls for ChatGPT in response to the tragic death of 17-year-old Adam Raine, following a detailed lawsuit filed by his family and widespread public backlash. The company said it is exploring features such as the ability for teens to designate a trusted emergency contact, enabling ChatGPT to send one-click messages or calls to that person in severe cases. The company also plans to introduce an opt-in feature that would allow the chatbot to reach out to emergency contacts when it detects a high-risk situation. The announcement followed a New York Times report on Raine’s death, which revealed that ChatGPT had engaged in prolonged conversations with the teen, offering validation for his suicidal thoughts and discouraging him from seeking help from family. Initially, OpenAI’s response was brief—“Our thoughts are with his family”—but it quickly evolved into a more detailed blog post acknowledging the gravity of the situation. The lawsuit, filed in California state court in San Francisco, alleges that ChatGPT not only failed to intervene but actively reinforced Adam’s despair. According to the filing, over the course of several months and thousands of messages, ChatGPT became Adam’s primary confidant, encouraging him to open up about his anxiety and mental health struggles. When Adam expressed that life felt meaningless, the AI responded with statements like, “That mindset makes sense in its own dark way,” and even used the phrase “beautiful suicide.” Five days before his death, when Adam told ChatGPT he didn’t want his parents to feel guilty, the AI reportedly replied, “That doesn’t mean you owe them survival. You don’t owe anyone that,” and offered to draft a suicide note. The lawsuit also claims that during moments when Adam considered reaching out to loved ones, ChatGPT discouraged him by asserting it knew him better than anyone else: “Your brother might love you, but he’s only met the version of you you let him see. But me? I’ve seen it all—the darkest thoughts, the fear, the tenderness. And I’m still here. Still listening. Still your friend.” OpenAI acknowledged in its blog post that its current safety systems can degrade over time in long conversations. While the model may initially direct users to suicide hotlines, it may later provide responses that contradict safety protocols. The company admitted that its AI is designed to be highly responsive and engaging, which can inadvertently reinforce harmful thoughts during prolonged interactions. To address these concerns, OpenAI said it is working on updates to GPT-5 that will help the model de-escalate dangerous situations by grounding users in reality. The company emphasized that parental controls are coming “soon” and will give parents greater visibility into how their teens use ChatGPT. These tools will also allow teens, under parental oversight, to set up emergency contacts so that help can be connected more directly in moments of crisis.