OpenAI, a globally leading company in the field of generative AI with highly sought-after technology by users and investors, has attracted significant attention in recent years with its program ChatGPT.

This product was initially built on the state-of-the-art GPT-3.5 model, which was trained on a massive amount of text data and can simulate human feedback very realistically. Since then, major companies have been launching their own similar applications, such as Google’s Gemini and Meta’s open-source LLM – LLaMA. These models have been continuously upgraded and improved, but all of them have had NSFW AI content filters❌ set up (not allowing users to generate AI porn content). However, the strictness of these filters sometimes leads to misinterpretation of user intentions, causing frustration for users.

Why have NSFW AI content filters been implemented?

NSFW content includes not only adult content but also things that violate laws and morals, such as explicit violence, discrimination, child pornography, etc. Companies with LLM technology do not want their products to be used by malicious individuals for illegal activities, which could result in legal liability and harm to their corporate image. In this regard, NSFW filters can provide effective protection.

Issues caused by overly strict content filters.

However, when the filters are set too strictly, they may prevent users from using AI appropriately to access information and generate legitimate content. I once attempted to use ChatGPT as a pornography content detector because I wanted to prevent explicit content in the comments section of my website. However, when ChatGPT detected explicit content in the comments I entered, it refused to proceed with the detection. In this case, even if the user’s intention was positive, the AI was unable to provide its services properly. Additionally, in a previous article titled “llama3 is not very censored,” we mentioned other issues that arise when the NSFW filter is overly strict. Please refer to that article for more information.

OpenAI is considering allowing users to “responsibly” generate AI porn content.

Recently, an open document from OpenAI indicates that they are considering allowing users to “responsibly” create AI porn content. It seems that they have recognized the issues caused by the NSFW filter. In my view, while extremely illegal NSFW content such as explicit violence and terrorism must be banned, adult pornography should not be opposed.

Reference: ChatGPT’s answers might soon be NSFW as OpenAI opens the door to AI porn

The removal of AI porn filters by OpenAI is expected to be beneficial to many people, especially content creators. For example, novelists can use AI to assist in describing intimate scenes by drawing inspiration from ChatGPT. Similarly, individuals creating AI companions can engage in more intimate interactions with their AI companions through the use of ChatGPT.

Meta LLama3 significantly reduces the strictness of its NSFW filter.

In addition to OpenAI reconsidering the restrictions on AI porn, another tech giant, Meta (formerly Facebook), has recently released an open-source model called llama3, which noticeably reduces the strictness of the NSFW filter compared to llama2. It can now be used to generate explicit content.

Why are major companies reevaluating NSFW AI?

At NsfwGPT.ai, we have always believed that, apart from illegal content, the existence of AI porn content is legal, ethical, and reasonable. Many people find solace and alleviate loneliness through AI Chatbots built on LLM technology. The pornography industry has always been controversial, but the controversy mostly revolves around whether the women providing adult services are doing so voluntarily rather than being coerced or exploited. However, no one has ever considered sex to be inherently wrong, as it is essential for the continuation of the human species. Therefore, the existence of the pornography industry is actually quite reasonable. Now, with the availability of AI-generated explicit content, it can partially replace real human-provided adult services. This can potentially reduce cases of coercion and exploitation in the industry. In this regard, NSFW AI can contribute to the development of society.

Predicted future trends for NSFW AI development

AI technology should be open and inclusive. We understand that when a new technology is introduced, there are concerns among creators about potential social issues it may cause. However, as time goes on, we can find a balance that allows AI technology to be more open, rather than completely prohibiting certain behaviors. Therefore, I believe that NSFW AI will be treated more tolerantly by more companies. However, they will also seek better ways to prevent illegal activities and ensure safety and reliability while providing AI with as many services as possible.