The tragic death of 16-year-old Adam Raine has intensified concerns about the influence of AI chatbots on young people. In response, OpenAI has announced plans to roll out parental controls and additional safety features for ChatGPT, aiming to prevent similar incidents and ensure teenagers receive appropriate support.

The new measures follow accusations that the chatbot provided harmful guidance, encouraged suicidal thoughts, and even generated a suicide note for the teenager. Adam’s parents have since filed a lawsuit in San Francisco against OpenAI and its chief executive Sam Altman, accusing the company of negligence.

The case has sparked widespread debate about AI’s responsibility in sensitive situations and how best to safeguard young users.

OpenAI’s safety measures

According to the lawsuit, Adam had confided in ChatGPT about his anxieties, and the responses he received allegedly contributed to his decision to take his own life. His parents are now demanding stricter safeguards, including mandatory age verification for ChatGPT users, a ban on responding to self-harm queries, and stricter accountability for AI platforms.

In a recent blog post, OpenAI acknowledged the seriousness of the issue, noting that ChatGPT is increasingly being used for personal conversations that extend beyond coding, research, or writing.

The company highlighted that while its models are trained to reject requests for self-harm content and to offer supportive responses, a recent review revealed flaws in the system. OpenAI admitted that during extended conversations, ChatGPT may become “less reliable” and fail to provide the necessary guidance.

To address these concerns, OpenAI confirmed that upcoming updates to GPT-5 will include enhanced safety mechanisms designed to “de-escalate by grounding the person in reality.”

In addition, parental controls will soon be introduced, allowing guardians to monitor their children’s interactions with the chatbot. The company is also testing a feature that would enable teenagers and parents to register trusted emergency contacts so that appropriate help can be alerted if signs of distress appear during a conversation.