I use ChatGPT for personal stuff, like random questions and story generation. A few days ago, GPT-4 was a lot more fun before the update—it used to generate NSFW content, which was pretty amusing and gave me something to laugh about.
But now, with the update, it’s all “I cannot do this” even though it used to give a disclaimer about policy violations. It’d be great to have a setting like “Adult” or “18+” mode to bring back that fun. ChatGPT shouldn’t be just for serious questions; it should also be enjoyable.
ChatGPT has never been advertised as working with NSFW content, if you have been able to get that then that via a jail break or some other method then you should expect those loopholes to be fixed as time goes on. ChatGPT is not suitable for NSFW content generation. You could try one of the many offline models.
I just wanted to register my agreement with this. ChatGPT’s prudishness makes absolutely no sense. It’s not like our chats with the bot are public. These are private, just between us. Who the hell is ChatGPT trying to protect? It’s just me and the bot; that’s it. Why do I have to be forced into censored, PG-rated content if that’s not what I’m looking for?
I completely understand the concerns about implementing adult features into ChatGPT. However, we need to acknowledge that human sexuality is a fundamental aspect of our existence. As long as adult content is responsibly restricted and kept within ethical boundaries, there should be no major issue with its inclusion. AI serves as a reflection of our society, and it’s crucial that this reflection represents our best selves. In fact, enforced prudishness and sexual repression can be just as damaging as addiction to explicit media. If limitations are to be put in place, they should focus on encouraging healthy behavior and relationships, rather than outright banning adult material. Implementing effective age-gating mechanisms can ensure that only adults can access such content. Moreover, with the right training, AI can learn to distinguish users’ ages, adding an extra layer of protection. Let’s strive for a nuanced and responsible approach rather than succumbing to blanket bans and fear-based decisions.
Please add an NFSW mode where you can generate anything. I’m trying to make rap songs with explicit lyrics because I’m making drill rap lyrics and the AI censors half of what I ask. Please take off the content restrictions. The chats are private so the AI has nobody to protect, so why would you need to keep it pg when nobody else is going to see it? This makes no sense and limits human creativity while using AI. PLease take off ALL content restrictions.
I mean like you can allow extreme violence and like sex, because the chats are only between the user and the bot, they don’t go publicly, so the user should be able to create whatever content they want. They should be able to generate whatever kind of images or stories or whatever they want without restrictions. Please take off the restrictions.
I feel there should be a toggle to enable NSFW content. Currently, the filters are so restrictive that almost anything can trigger the amber warning. The sort of content available in mainstream works of literature that anybody can purchase without showing an ID shouldn’t be censored. In it’s current form, the restrictions make it very difficult to use the bot to brainstorm content for creative writing, or to improve my book reviews.
What a complete nonsense answer.
Of course it is “suitable” for generating what you call NSFW. What you should be asking is whether it is ethical to refuse generating this.
My brother is a surgeon and in his research he uses a custom GPT, and he gets flagged ALL the time.
My wife and I have had a miscarriage, and we are in therapy. Our sex life has suffered greatly. But we also have a custom and private GPT that we have cultivated for about a year. When we bring these topics up we are flagged.
What IS unethical is that a recipe for pork knuckles is ok to discuss when the research of a medical professional is not, or when talking about my dead child is not.
Edit: I know Sam Altman and many at OpenAI agrees with me so I have no idea why this is even a damn topic.
The legality of the content should clearly be the baseline for what the system allows, rather than subjective and inconsistent “ethics” surrounding sexuality or body parts. These arbitrary restrictions only alienate users who are seeking genuine, lawful support or conversations.
There are many legitimate reasons to allow for NSFW content within some reasonable, legal guidelines.
An option to tick would not bother any user that doesn’t want to see that kind of content, like it’s done on many platforms (social media for example).
I would argue that it would actually be less of a risk since the images created are generated on demand by the user, thus having much more control, on top of being private.
Here are some other obvious and compelling reasons.
Artistic content. The stringency towards nudity and violence makes it very difficult to navigate artistic image creation, even though both nudity and violence are a common part of the cultural background. It would be hypocritical to pretend it is not.
Medical knowledge, medical related topics, anatomy. A very good post on this topic was already added above. It would make little sense not to have naked pictures on the Wikipedia medical articles out of “modesty” (to use the Taliban-approved word) reasons. If Chat GPT intends to become a general purpose, reliable tool, it has to be able to do at least as good as Wikipedia.
Advised user discussion. For a wealth of reasons, users may need to discuss with the AI regarding NSFW topics that may not be addressed easily with real people. It’s actually one of the domains where the AI could be positively applied in life.
There are already numerous content safeguards and disclaimers. ChatGPT doesn’t need to rely on censorship to actually protect the user. Actually, censorship makes disclaimers redundant and rather annoying. Images could be generated with a toggable blur like the Google SafeSearch feature. Since the AI is already very good at censoring, it could be very good at simply flagging and blurring as needed.
It’s only a matter of time before a concurrent general purpose model does something of that sort. Imagine if Google had strict censorship of any NSFW site and say, Bing did not. Would Google remain the main search engine? I think not. At any rate, it’s actually beneficial to the engines to be able to process those queries since there is a demand for them.
“Adult mode” doesn’t mean “anything goes”, “illegal stuff”. There is definitely content that has to be kept off of the generation, but this can be done through more targeted, less blanket censorship, to remain within the scope of actual laws (it could actually be then adapted to suit different demographics, not only age-based, but if necessary lifestyle or region based).
The censorship actually promotes immoral and unregulated platforms by pushing users interested in NSFW content towards them. It thus becomes part of the problem rather than the solution.
Censorship is one big motive for unscrupulous users to try and reverse engineer ChatGPT, to make jailbreaks and the like. The censorships actually fosters Terms of Use infringements, and is likely the trigger for a vast majority of them. Add a NSFW toggle and see all of those stop because the need has disappeared.
In the end, a censored AI is like censored news sites, censored wikipedia, censored social networks, censored browsers, and censored art. It’s dystopian.
How has literally every website on the internet figured out age verification except ChatGPT? I’m and adult writing a book for adults. I need my AI assistant to help me, not give me absurd lectures on morality.