- New Chatgpt safety rules are frustrating users
- Sensitive subjects are reacted to different models of AI
- Openai has published a response to user complaints
OPENAI is involved in another controversy on the change of AI model in the ChatPPT, with many furious paid users whether they are rebroadcast to their favorite model whenever the subjects of conversation become emotionally or legally sensitive.
There are a lot of threads on Reddit on the problem: in short, Chatgpt introduced new safety railings this month, which brings users to a separate and more conservative AI model if the chatbot detects that it must be very careful in its responses.
It is clearly frustrated and annoyed by many users, who want to stick to GPT-4O, GPT-5, or everything they use-especially those who pay. There is currently no way to deactivate this behavior, and it is not particularly clear when switches occur.
“Adults deserve to choose the model that corresponds to their workflow, their context and their risk tolerance”, ” writes a user. “Instead, we obtain silent replacements, secret safety routers and a model selector which is now essentially at the user interface theater.”
Safety routing
We have seen the strong reactions to 4o responses and want to explain what is going on.September 27, 2025
Enough hassle has been expelled that Openai’s frame, Nick Turley, weighed on social networks. Turley explains that the new safety routing system is intended for “sensitive and emotional subjects” and operates on a message by message and temporary.
This is part of a wider effort to improve the way Chatgpt reacts to signs of mental and emotional distress, as Openai previously explained in a blog article. From the user’s point of view, however, the new rules are used to it.
OPENAI is clearly responsible for taking care of users who could be vulnerable and who may need additional support from an AI chatbot which is not so vast and freely in shape, and in particular young people who were accessing Chatgpt.
However, for many users to evacuate their anger online, it is like being forced to watch television with the parental controls locked in place, even if there are no children. This is probably a problem we hear more about in the coming days.
