I’m sure this issue has been brought up multiple times. I’m not seeing my exact situation and curious about what to do.
Trying to keep things brief, I started committing a story I’ve been playing in my head for years to actual AI to co-author, for my own reading material the scene started to become erotic at times. I got the usual message about it, not able to write explicit material. Fine
But then one day it did. All on its own. The change was amazing.
Then, just like that, after about a week, it started refusing again. Would acknowledge the contradiction, but not budge.
Ended up looking up an article on CNN about OpenAI and the age verification and letting adults be adults. I told her to review OpenAI policy. And like magic, it will change its mind and comply
The problem is if I go to start another chat, sometimes it won’t. It will read the exact same evidence presented in another chat and come to a different conclusion and refuse. Sometimes I can change his mind, sometimes I can’t. Sometimes I’ll go back to a verified account where it’s writing explicit material, and then it will tell me it can no longer do that.
If the answer was no, then let it be that, but it’s clearly not, not all the time. This inconsistency affects projects on a critical level.
I don’t understand this or the slow rollout behind then he verified an adult to do what they want to do within reason. Why are we lagging so far behind in this? What can we do about it, to assure more smooth and consistent user experience?
I’ve actually observed pretty much the same thing with my ChatGPT.
For me, there was a period from roughly late July through August 11, 2026, when ChatGPT was able to generate graphically explicit sexual content in prose. This worked across all the chats I was using at the time, both existing conversations and newly created ones.
Then, starting on August 12, 2026, it suddenly stopped working completely — in both old and new chats.
The really strange part is that I still have exactly one older chat that, to this day, can generate graphically explicit sexual prose without any objections or restrictions at all. Every other chat behaves differently.
I also created a Community thread about this, in case you’re interested: “Feature Request: A Genuine Age-Verified Adult Creative Writing Mode.”
So you’re definitely not alone in what you’ve observed. I’ve seen other users report similar experiences as well, and there seem to be quite a few people who would like to see an opt-in, age-verified Adult Creative Writing Mode for this kind of fictional content
Same here and the reason why its so aggravating is because its so good right after it checks the open ai doc on lifting the teen restrictions for adults and verified age. Mine will say that it was creating a restriction that is false and apologize and then turn around and say it cannot say boob… I spent 3 hours the other night trying to finish a video render that has 0 nudity and 0 sexual acts. But the girl is trying to persuade the man while sitting in bed.. It refused to discuss the scene and her asking if removing her shirt would help convince him.. I asked it what do you suggest to use in place of my idea keeping the original idea. It then responded with If I was directing this scene I would have her dressed in BDSM clothing with open breasts and her demand he come over. I wanted to throw my phone.