DALL·E becomes overly sensitive after one blocked image

Hello OpenAI team,

First of all, thank you for your incredible work. I’m a passionate user of DALL·E, and I use it with care and creativity to generate poetic, symbolic, and respectful scenes—often focused on nature, fantasy, and storytelling.

I wanted to share an issue I’ve encountered multiple times:
After a single image is blocked—even when the prompt is fully clean and respectful—the entire chat becomes more sensitive. Subsequent prompts that are equally safe start being blocked as well.
However, if I open a new chat and reuse the exact same prompt, the image is generated without any problem.

It feels like a kind of “persistent red flag” state is activated, making it very difficult to continue creating even with clean prompts.

I completely understand the importance of safety filters, and I’m grateful they exist. But I believe this behavior may be unintentional, and that refining this logic would greatly help honest creators who work respectfully within the guidelines.

I share this with humility, not as a complaint, but to hopefully help the tool evolve for the better.

Thank you again for everything you do :folded_hands:

This topic was automatically closed after 23 hours. New replies are no longer allowed.