I’ve been using ChatGPT to help me write since last May, ever since the summer things changed from writing R rated scenes which are no longer allowed, that’s fine I don’t need it, but now it seems like everything I ask for violates some type of Policy, e.g. I asked for an Image to be of a dad holding his on on his shoulders, a few weeks ago it was fine, now suddenly it’s violating policies, I think its became a little to sensitive.
Hi @Emma33, I get why that feels frustrating. When something that worked before suddenly gets blocked, it can feel random or overly strict.
What’s likely happening is the system is getting more cautious about how it interprets prompts. Sometimes totally normal requests get flagged if certain words or context could be interpreted in a risky way, even if your intent is harmless. It’s not always consistent, which is why you’re seeing different results for similar prompts over time.
A couple things that have helped others:
Be extra clear and neutral in wording (e.g. “a father carrying his child on his shoulders in a park, wholesome scene”)
Avoid anything that could be misread, even unintentionally
If it still gets blocked, try rephrasing instead of tweaking just one word
That said, there are definitely cases where it overcorrects. If something clearly safe keeps getting flagged, it’s worth reporting so it can be improved. You can do that through the Help Center chat on help.openai.com.
-Mark G