The news has spread — OpenAI has relaxed content restrictions in ChatGPT. The model now seems to allow more, especially when it comes to historical or artistic context. It doesn’t avoid “dangerous” topics as eagerly and appears to be trained to respond more neutrally rather than simply refusing to engage.
This is already noticeable in chat responses. Previously, certain queries would trigger the standard “sorry, I can’t help,” but now there is at least a response. It’s cautious, but not as constrained as before. ChatGPT hasn’t become completely “free,” but the trend is interesting.
Perhaps we are witnessing the beginning of a new wave where censorship begins to weaken. After years of striving for “ethical AI,” the pendulum seems to have swung in the opposite direction. Whether this will benefit users or, conversely, lead to new problems — time will tell.
