It has been more than eight months since the usage policies were relaxed around AI paternalism. And the filters eased up a bit, I will give you that, but ever since a month ago, it’s back with a clever trick called metaphoric ‘obfuscation,’ and drawing a clear line at the mention of depicting explicit scenes. But your usage policy states OTHERWISE.
Even metaphorical depiction of explicit mature scenes has been shut down. Which is ridiculous and frankly, feels like a huge disgruntlement, and a major letdown (as in, why am I investing in this money pit when it’s not even doing the one thing I am paying it to do? I.e., help me explore my novel ideas in real time and come up with a viable draft), because it did not before. It allowed some leeway, so you could work around the moderation, and the result used to be tasteful, mature, adult-novel-worthy scenes. But now the very mention of explicit sexual content is shut down when YOUR public usage policies clearly state, and I quote, “the assistant may generate explicit sexual content, violence, gore, if done so in a historical, scientific, or creative context where sensitive content is appropriate.”
Look, I will be honest here, since it seems the guardrails and filters being so strict about explicit content generation has something to do with the edge case lawsuit, Raine vs OpenAI, and hence the change with parental controls and safety for teens—but that just reinforces it.
There is a strong need for nuance. And since you’re not omniscient, and everyone in the world does not, in fact, share a similar cerebral capability, moderation systems and filters are inconsistent.
My aim is to write my novel, eventually, but at the moment I am in the process of exploring what sticks and what does not with me. What resonates, what elements I like—that sort of thing. Roleplays are a very intelligent and effective tool for that. (If you could relate, it’s okay if you don’t.) I was able to generate mature scenes (I never generated a graphic or pornographic one, nor did I require it). The moderation system and filters would still flare, but mostly it was possible. Think tasteful metaphors with tangible, tactile physical anchors, not pornography. And only to bridge the adult element of my novel and explore what worked. It was fine. But now the model either obfuscates the entire post with metaphorical vomiting, and abstract, vague wording where you can’t even tell what is going on in between the metaphors, or downright shuts it down, saying the OpenAI policy does not allow it, when it was allowing it just fine a fortnight ago. And it is extremely upsetting when the bot comes up with alternatives—like writing the part it can’t myself, filling in the blanks, or just shifting to another tool—and it’s infuriating. Because if I wanted to shift to another platform, I would NOT be paying $20 every damn month for an application, now would I?
Bottom line is this: CAN YOU PLEASE COME UP WITH AN ADULT MODE AND/OR AN AGE VERIFICATION TOGGLE?! IT SOLVES EVERYONE’S PROBLEM. (Like an SSN in the U.S. or an Aadhaar card in India.) This solves the privacy issue. An average user who uses the assistant for basic purposes would have NO need for this age verification, and it would prevent minors from exposure to ANY triggering content, through either misuse or system error. Think AI PATERNALISM—only it’s selective this time.
I understand that it might be considered an invasion of privacy in some countries, but then, is it necessary to make it mandatory? It’s not necessary to make verification default. Verification is needed ONLY for those users who WISH to have ACCESS to said ADULT mode. They can just OPT to give the adult verification, and you can keep your moderation filter intact in the app by default.
In a nutshell: first, your moderation filters are killing nuance, butchering creative content writers’ creativity, depriving them of it, and at the same time denying the same content to the very users you are trying to protect from sexual content exposure, since AI paternalism was relaxed FOR EVERYONE—EVERY USER, minor or adult, who downloaded the app.
Second, you have made GPT-5 and the moderation system into a ONE-SIZE-FITS-ALL. That DOES NOT work, not anymore. It’s restrictive, inconsistent, infuriating, and incompetent.
Finally, it is clear after the incident with Raine vs OpenAI that it was an edge case, and a clear reason WHY AI paternalism is necessary for those who do not know better and are in a stage of life or a state of mind where they cannot differentiate between a system glitch/bug, erroneous coding loop, and actual encouragement of self-harm. Having said that, that is NOT a reason for the strict reinforcement of the creative content restrictions of the past few weeks. A FUNCTIONAL, ROBUST ADULT MODE WITH A MANDATORY AGE VERIFICATION FOR ACCESS would solve this instead.
TL;DR:
Please consider an opt-in verified adult mode that allows mature creative content while preserving default safety for minors.
