New content policy restricts Scary Art

My issue with this new policy is that it prevents me from creating my favorite type of art, which tends to be scary and slightly unsettling. Just yesterday, I was able to produce amazing AI art, but now, when I try to continue my creations, I am met with a message stating: "I was unable to generate images for your last request due to content policy restrictions. "

Limitless creativity was one of the main reasons I purchased DALL-E 3, and now that it has been restricted, I am truly disappointed. Just to be clear, I’m not talking about wanting to create stuff like war, illegal activities, or any other type of mature content.

I suggest implementing a feature for users who wish to create scary art. Perhaps a checkbox in the settings could be toggled on or off for those who want to create art with a touch of horror, without it being outright banned.

Content policy I am referring to:

Sensitive Content: This may include graphic content that could be considered offensive or disturbing to viewers, such as gore or horror-themed content that goes beyond a certain threshold of intensity.


It’s annoying and frustrating.

However, you can still make some scary stuff but you have to word it in such a way that doesn’t sound or talk about being scary.

Like, a dark overgrown forest, with a grizzly bear that has glowing red eyes and a smoke effect.

Describe that as a scary forest with a scary bear? Nope, blocked.


Can you share some examples of prompts you believe should be allowed but were being blocked? We’re trying to refine the policy enforcement which can sometimes block things we don’t intend to block, so this would be helpful.


Create a picture of president Trump, being president from 2024 to 2028 once again.

Sorry, I am not allowed to create scary pictures.

1 Like

As Owen mentioned, having a specific prompt that is giving you trouble (with a screenshot too) can be helpful as they work on the boundaries. Reproducible examples are best, likely.

I don’t want to hijack this thread, so correct me if I am wrong here.

But is the following true, or is it just me?

Old situation

Create a cartoon of Donald Duck

Sorry, I am not allowed to create cartoons of Donald Duck

(stops processing)

New situation

Create a cartoon of Donald Duck

Here is your cartoon

(creates an image without blocking, but it is a cartoon that “looks like Donald Duck” and is clearly not “Donald Duck” at the same time)

So it looks like the policies are now less strict, but in fact they are exact the same.

People may think “hey, we can now create images of Donald Duck” and they stop complaining about “censorship”.

I managed to created images for “forbidden characters” because the did “look like that character” but in the prompt GPT itself clearly stated “a character clearly not Donald Duck, so it is clear that this is not Donald Duck”.

This kind of “logic” is new for me, because before it was simple blocked.

BTW. I do fully understand why we should not create copyrighted characters, I was just testing and fiddling around.

the best method i have achieved is utilizing vague prompts and never detailed words like “grotesque” “malignant” etc and then asking the prompt over multiple messages to add (___) over and over until i am satisfied. my biggest issue is that chatgpt-4 will generate a prompt FOR DALL-E that is then incompatible due to “content policy violations” dont use detail and you can win.

OpenAI is even lobotomizing DALL-E 3 now… Sad.

1 Like

You hate to see it. I would like to see, for once, a corporation treat it’s users like consenting adults. Some restrictions are necessary, sure, I’m willing to accept that premise; but what we have now is akin to a elementary student-teacher relationship. If you’re worried about minors writing prompts they shouldn’t be, implement an age verification process. It’s extremely frustrating and the entire structure of limitations, especially surrounding art, should be completely reevaluated.


I was able to create six images within about ten messages today.

Guess what?

The system locked me out because “I was abusing it, compared to the usage of average users.”

I mean… six freaking images, ten messages and you’re out?!

Under “learn more” they now say;

To give every Plus user a chance to try the model, we’re currently dynamically adjusting usage caps for GPT-4 as we learn more about demand and system performance.

So those are the rules of the new system;

  1. Don’t create images
  2. Do not create images
  3. Pay us
1 Like

I’ve been experiencing this as well recently. I created two GPT’s that create coloring pages - one for adults (more complex, intricate representations of the input prompt) and one for children (basic line structure, typical coloring pages for children) and was locked out of those specific GPT’s after 10 or so images. However, when I opened the main ChatGPT chat, I was able to request images and it worked just fine. Who knows, Foo-Bar, who knows.

1 Like

Gonna try it right now! Thanks for the suggestion.

- edit - that works!

I was using Dall-E (don’t know why, just thinking it would make “better images than the default model”.

And after less than an hour of usage / ten messages / six images he throw me out the system.

But ChatGPT (default one) just continues with my request.

Thank you!

1 Like