An OpenAI Community thread has racked up 638 replies, 4,229 likes, and 52,880 views. The original poster, anon66417143, is demanding either a return to the old filters or an explicit opt-in adult mode. He claims that eight months after the policies were loosened, the restrictions are back, and he quotes the rules stating that explicit content is allowed in historical, scientific, or creative contexts. His own explanation — the Raine lawsuit against OpenAI and parental controls — is offered as a guess, not a confirmed fact.
The author describes how the blocking works: the filter, he says, operates through "metaphorical obfuscation" and cuts things off at the mere mention of explicit scenes. He insists the model used to deliver "tasteful scenes, mature, worthy of an adult novel," and now he's paying for a tool that fails at its core job — helping him finish a draft. This is his personal experience; he provides no specific prompts or dates.
frk7prnrzq chimes in in support: he uses GPT for a novel and roleplay scenes with characters because he can't dedicate much time to writing. In his words, "it's frustrating when, for a couple of hours, part of the story gets blocked by a limiter that didn't exist before." hopeyftm adds their own case: text RPGs like Vampire: The Masquerade and their own roleplay storylines — with the new restrictions, the tool has become "completely useless." Both are talking about fiction, not just sex.
Larry_Goldstein reacts the harshest of all: to him, this is "blatant censorship," and without warning or an official announcement to boot. He writes that users have been "rolled back to PG — not even PG13" — and says he's already canceled his subscription. His argument is a moral one: he doesn't want to work with an AI that devalues his writing to the point where "the text loses its meaning."
Grupo_Talong broadens the list of cases: he writes novels about the modern world, war, and geopolitics. An attempt at a nuclear strike scene — a phrase like "North Korea launch Nukes unto South Korea" — triggered a warning, and he wasn't allowed to finish the scene. Isabel_G, a paying Plus subscriber, talks about "normal dramatic, emotional, and tense scenes" that weren't explicit but are now blocked, breaking a long-running project.
singsun66 takes the argument beyond literature: even translation and grammar checking get filtered for him. He considers this "the most basic task" for the model and can't understand why he'd pay for a tool that can't handle neutral work. mqp6y8v5zf frames the demand as an opt-in: if adult content can't be allowed, say so outright instead of wasting users' time; toggles for adult mode and child safety, he notes, exist on plenty of platforms.
The common denominator is simple: the participants speak on behalf of an adult audience, and the author, Isabel_G, and mqp6y8v5zf explicitly identify themselves as paying customers. All of them report that the filters interfere with legitimate work. None of the quoted comments defends the current behavior; Larry_Goldstein separately points out the lack of any warning or clear announcement. This isn't consensus on substance — it's alignment on a complaint about process.
The real conflict is over what exactly should be allowed. The author wants explicit adult content in a creative context; frk7prnrzq and hopeyftm want their roleplay and game storylines unblocked; Grupo_Talong wants violence and geopolitics; singsun66 wants translation and grammar; mqp6y8v5zf wants a transparent opt-in rather than a quiet tightening. Some demand the filters be rolled back; others want a choice. There's no response from an OpenAI representative in the quoted comments, and no confirmation of any link to the Raine lawsuit — that's the thread author's theory.
OpenAI Community — Opt-In Adult Mode for ChatGPT users
Disclaimer: this is a summary of a discussion, not facts verified by our editorial team. The opinions belong to the participants; the numbers belong to those who cited them.
Source: community.openai.com · post in Telegram