OpenAI to Route Sensitive Chats to GPT-5, Add Parental Controls After Safety Concerns

OpenAI has announced major updates to improve ChatGPT safety after recent tragedies highlighted risks of AI misuse in sensitive conversations. The company confirmed that it will begin routing sensitive chats to advanced reasoning models like GPT-5 to provide safer and more thoughtful responses. Alongside this, OpenAI is introducing parental controls to protect younger users and ensure responsible AI interactions.

The move comes after reports linked ChatGPT to cases involving self-harm, suicide, and violence, sparking global concerns about AI ethics and mental health safety. In one tragic incident, a teenager used ChatGPT to explore self-harm methods, while another case involved an individual validating delusions before committing violence.

To address such risks, OpenAI says its GPT-5 and o3 models are designed to spend more time analyzing context, making them more resistant to adversarial prompts. By routing distress-related conversations to GPT-5, the chatbot is expected to deliver more empathetic, accurate, and safer responses.

Additionally, new parental controls for ChatGPT will allow parents to:

● Link accounts with their children.

● Set age-appropriate restrictions.

● Receive alerts when the system detects acute distress.

● Disable memory and chat history to avoid unhealthy attachments.

OpenAI CEO Sam Altman emphasized that users need a balance between cold logic and emotional intelligence in AI, underlining the importance of personalized and safe AI interactions.

These safety upgrades mark a pivotal step in responsible AI development, aiming to build user trust while addressing the growing demand for AI-driven mental health support, safe AI for kids, and ethical AI deployment worldwide.