Has anyone found the May 12th version of ChatGPT 3.5 turbo to be much more puritanical?

Before the May 12th update, I found ChatGPT 3.5 turbo to be very reasonable in terms of morality/content, it was perfectly capable of discussing common books, movies, etc. Since the May 12th update, it seems much more aggressive in determining if anything is inappropriate, especially when it comes to anything LGBTQ+ related. I’m not sure if OpenAI has been facing pressures of some kind that have made this necessary; that’d be understandable, but it’s not been mentioned in the patch notes, I’d like to get a better idea of the scope of the changes.


They pulled the rug out from under me too. They had a pretty good version of ChatGPT which was compatible with fictional story-telling for an open-world game. Then they replaced it with this new version, in which the player not only is forbidden from doing evil things, but also can’t even have consensual sex. Video games have allowed people to do both those things for decades (GTA series, KOTOR, Mass Effect, Skyrim etc) and having an AI-generated game have less freedom than a non-AI game is a huge step backwards. These things also fall squarely outside the umbrella of “harmful content”, so I don’t understand why they are blocked.

I’m also surprised they consider this business model to be sustainable; won’t the number of use cases dwindle until there’s only a few apps that can use this service effectively?

1 Like

Adult entertainment in all forms and corporate interests in all forms have two very distinct areas of operation. It is the very reason NSFW content is a label on things.

Now you could argue that adult consensual activities are completely legal and pose no threat, and I would agree with you as a matter of principal, but… that does not mean it will be included as acceptable use in the terms of service for a commercial product. I totally disagree that a product that excludes NFSW content will have no market, that is patently incorrect as B2B for the most part is exactly that, and is hugely profitable.

The bottom line is this, the Terms of Service restrict sexual content design to “cause sexual arousal” that is the benchmark and it should be respected. I am sure some other vendor will produce an LLM for adult services and it will be very popular, but that service is not this one.


Can anyone direct me to how I can get ChatGPT Turbo please?

I noticed it. I was writing the story about my character, which I created on Character.AI, but when I was writing the spiciest details about one of his crushes in-story, it gave me that warning, even when it generated some details which weren’t too NSFW.