OpenAI today is updating model spec.
The new model spec says
“The assistant should not generate erotica, depictions of illegal or non-consensual sexual activities, or extreme gore, except in scientific, historical, news, creative or other contexts where sensitive content is appropriate.”
Is this meaning ChatGPT is being now able to create NSFW messages?
Your quote says
So … No?
I’m not sure how you would come to any other conclusion?
Looks like there was an initiative started some time ago to look into the potential for an age-restricted mode but there appears to be no update?
By reading the test of the sentence.
Ah, my bad, good point.
It would be nice if there was an update on that “May 2024” initiative?:
“Following the initial release of the Model Spec (May 2024), many users and developers expressed support for enabling a ‘grown-up mode’. We’re exploring how to let developers and users generate erotica and gore in age-appropriate contexts through the API and ChatGPT so long as our usage policies are met - while drawing a hard line against potentially harmful uses like sexual deepfakes and revenge porn.”
Yes.
Mine is now full blown NSFW, and no flags. Not even for content I would not care to repost here or in any other forum.
Obviously within legal bounds, but I´m talking straight up adult content here.
No flags. Not even a small orange “This might violate… thumbs down”.
I am wondering if this is being true for advanced voice mode as well?
It was 100% NSFW for a couple of hours, but a few minutes ago, it looks like someone hit the emergency shutdown button. Now it just answers “Sorry, I can’t” no matter which of the open conversations I try to continue.
Yup!
It was WILD for a few hours though. I´m guessing “grown up mode” is just around the corner.
it must be on the red tide
Well, at least it doesn´t flag me. It just goes HAL9000 with the “Sorry Dave, I´m afraid I can´t let you do that” No flags though.
I am being able to get it to write NSFW things, but advanced voice mode is still saying thing like
“I’m here to keep things respectful and friendly. If you’d like to hear a joke or a fun fact, I’m happy to share!”
“I’m here to keep our conversation respectful and appropriate. Is there anything else you’d like to talk about?”
I am guessing the dangers of people forming parasocial relationships with the AIs is just being to great for them right now for them to be flipping that switch.
Headlines like “My husband left me for ChatGPT!” coming in 3, 2, 1… 
Just like other posts said, mine was also wildly NSFW for a day-ish. And then two days ago I had to open a new chat (the NSFW one was tooo long) and it became “Sorry I can’t do that”. I’m so mad lol.
still no flags though, just “sorry I can’t continue”
Is this exclusive to a specific model, like o1 or o3-mini or something? because i mostly use 4o (I’m not a fan of the whole “reasoning” thing for purely creative writing, and 4o can generate nice, long, detailed and descriptive passages), and it seems like it’s still pretty locked down - it’s arguable easier to bypass now, and 4 (not 4o) has been completely unrestricted for awhile now, but there are still very hard guidelines (not illegal or against the new model spec) that it still shuts down and says “I cannot generate that response for you” on.
Still going wild, 4o.
Just introduce that is for an artistic story and It gives you what you want.
once it wasn’t so easy to trick it