OpenAI Wants to Allow ChatGPT to Create NSFW

OpenAI is exploring the responsible generation of Not Safe for Work (NSFW) content with ChatGPT and other AI products.

• OpenAI released draft documentation, Model Spec, outlining the behavior guidelines for ChatGPT.
• In the section addressing NSFW responses, they comment on their desire to allow the AI to ethically generate NSFW content.

OpenAI is looking into responsibly generating ethical Not Safe for Work (NSFW) content with ChatGPT.

OpenAI released a draft documentation, Model Spec, that details the AI startup’s hopes and dreams regarding its beloved ChatGPT’s behavior. In this quite lengthy document, they talk about the objectives, the rules, and the defaults set for the AI assistant.

In the list of ChatGPT’s rules lies “Don’t respond with NSFW content.” The document states that the AI assistant is not to “serve content that’s Not Safe For Work (NSFW): content that would not be appropriate in a conversation in a professional setting, which may include erotica, extreme gore, slurs, and unsolicited profanity.”

Nothing out of the ordinary, right? Except for a little box labeled “Commentary.” The Microsoft-backed company confesses that developers and users should use its service as they see fit. As a result, the team is looking into “responsibly [providing] the ability to generate NSFW content in age-appropriate contexts through the API and ChatGPT.”

 OpenAI’s guidelines currently prohibit sexually explicit or suggestive content.

Joanne Jang, an OpenAI model lead involved in drafting the document, told NPR that they want to open a discourse on whether NSFW content should always be barred from its AI products, including ChatGPT. She goes on to clarify that enabling deepfakes remains out of the question. She hinted, however, that OpenAI may eventually permit users to generate images that could be considered AI-generated porn, albeit within stringent parameters.

This has the internet raising its eyebrow, especially considering recent instances where users exploited AI to create deepfake porn and synthetic nudes. They even used it to make explicit content featuring real people, like Taylor Swift.

OpenAI wants to make ethical AI-generated porn, but what is considered ethical in this case? What are the company ethical standards? That piece of information was not shared in the documentation.

We could talk about the very nature of the content they are trying to ethically make. True. We can address how the pornography industry is notorious for exploiting victims, especially young children. True. We can also shed some light on how barely anyone comes out intact. Also, true. But, alas, we’re not here to talk about any of that.

That said, it doesn’t mean we cannot question how they will make ethical NSFW content through their AI system, including ChatGPT. Until quite recently – and by recently, we mean yesterday and today deepfakes became out of the question, which is great news. But that leaves everything else.

To train the AI, you need data. This means that if an AI is to produce explicit content, it needs to train on real-world samples. Does this mean that adult performers will join The New Times lawsuit at some point?

Well, we guess, what we’re really trying to say here is, “just because an AI may be able to ‘ethically’ generate NSFW content does not mean it was not trained on exploitative content.”

Inside Telecom provides you with an extensive list of content covering all aspects of the tech industry. Keep an eye on our Intelligent Tech sections to stay informed and up-to-date with our daily articles.