OpenAI rolls out safety routing system, parental controls on ChatGPT 

OpenAI begins a new test The protection routing system In the ChatGPT on the weekend, and on Monday introduces parents controls in the chatboat – drawing mixed reactions from users.

Protection features come in response to numerous events of specific Chatzipt models that make users misleading thoughts instead of rebuilding conversations. Opena is the face of a Wrong death suit A teenage boy is bound to a such incident after a few months of conversation with ChatzPT after suicide.

The routing system is designed to identify emotional sensitive conversations and automatically switch to Mid-Chat GPT-5-thought, which the company looks as the best equipped model for high-distance protection work. Specifically, was trained with GPT -5 models A new security feature that calls openly “safe perfection”, “ Which lets them answer sensitive questions instead of just refusing to be involved.

This is a contrast to the company’s previous chat models, which are designed to be compliant and answer the question quickly. GPT -4O is under special investigation because of the extra psychopantic, consent nature, which both have drawn a large base of fuel and dedicated users to the AI ​​-inflammatory confusion. OpenAI when rolling GPT -5 as default in AugustMany users have demanded access to GPT -4O by pushing behind.

Although many experts and users welcome security features, others have criticized what they see as an additional caution, some users have been accused of treating the opening of adults in such a way that reduces the quality of service. Opeena suggested that it would take time to get properly and gave yourself a 120 -day repeat and improvement time.

The ChatzPT VP and chief Nick Tarley explained the router with the implementation of the router with the explanation of “Strong reaction to 4O reactions”.

“If routing occurs on the basis of per–message; switching from default model occurs on temporary basis,” Turley Posts PostThe “Chatzipt will let you know which model is active if asked. It is part of a broad effort to strengthen the protection before a greater rollout and to use the real-world” ”

Has received implementing parents’ control in ChatzPT Similar level of praise and condemnSome praise parents to give their children a tab in the use of AI, and others are afraid that it opens the door to treat adults like the opeina kid.

The controls set the parents calm time, shutting down voice mode and memories, removing the image generation and out of model training allows them to customize their teenager’s experience. Will also get adolescent accounts Extra content protection -Shras graphic content and extreme beauty ideals-and an identification system that recognizes potential symptoms that a teen may think about self-loss.

“If our systems detect potential losses, a small group of specially trained people review the situation,” per Open EEE BlogThe “If there are signs of intense crisis, we will contact parents through the text, text message and be careful on their phone, unless they choose.”

Openly acknowledged that the system would not be perfect and sometimes the alarm could be extended if there was no real danger, “but we think it is better to act and warn a parent so that they can take steps rather than remain silent.” AI Farm says it is also working on law enforcement or emergency services that it is also working if it detects an imminent threat for life and cannot reach the parent.