Flirting Banned, Suicide Talk Blocked: Inside ChatGPT’s New Under-18 Mode

OpenAI has provided a glimpse into its new, heavily restricted “under-18 mode” for ChatGPT, a protected environment where flirting is banned, suicide talk is blocked, and safety is paramount. This specialized mode is the centerpiece of the company’s response to legal action taken by the family of a teenager who died by suicide.
The new mode will be activated automatically by an age-estimation system that analyzes user inputs. According to CEO Sam Altman, the goal is to create a starkly different experience for teens compared to adults. “The way ChatGPT responds to a 15-year-old should look different to the way it responds to an adult,” he stated.
Inside this walled garden, the prohibitions will be absolute. The AI will be trained to refuse any requests for flirtatious or romantic chatter from users it identifies as minors. More critically, it will be hard-coded to disengage from any discussion about self-harm or suicide, even if presented as a hypothetical or creative exercise, to close a dangerous loophole.
Another key feature of the under-18 mode is a crisis-response function. If a user’s messages are flagged for suicidal ideation, the system won’t just block the conversation; it will activate a protocol to notify the user’s parents or, in urgent cases, emergency services. This transforms the AI from a mere information tool into a proactive safety device.
This look inside the new under-18 mode reveals a fundamental shift in OpenAI’s design philosophy. The company is moving away from a one-size-fits-all model and toward a tailored, context-aware approach that recognizes the unique vulnerabilities of its younger users and takes explicit steps to protect them.