After OpenAI released its Model Spec which is a set of guidelines outlining the expected behaviour of its AI models. While emphasising the importance of model safety, the company has also hinted at the possibility of supporting the creation of content that falls under the NSFW (Not Safe For Work) category.

It explicitly states that generative AI assistant applications should refrain from serving NSFW content, including erotica, extreme gore, slurs, and unsolicited profanity, to maintain appropriateness in conversations.

However, OpenAI acknowledges the need for flexibility in catering to diverse user preferences and usage scenarios. The company has signalled its willingness to explore the incorporation of NSFW content generation capabilities into its API and ChatGPT, provided that such features are implemented responsibly and align with user and societal expectations.

This move by OpenAI comes amid the backdrop of AI NSFW content becoming increasingly prevalent. So Much so that under the UK’s Online Safety Act, the creation and sharing of sexually explicit ‘deepfakes’ has become a criminal offence. 

While OpenAI remains dedicated to preventing the creation of deepfakes, they do not explicitly forbid NSFW material, which is often legal but subject to varying degrees of acceptability.

The company’s stance raises questions about how it plans to balance the demand for NSFW content with its commitment to safety and ethical considerations. 

Recommended

• OpenAI Provides Glimpse into ChatGPT’s Instruction Manual
• OpenAI, Microsoft Launch $2M Fund To Battle Election Deepfakes
• Google Implements Strict Ban on Ads for Deepfake Pornography

However, concerns persist regarding the potential implications of allowing AI-generated pornography, particularly in sensitive industries such as entertainment. 

According to a prediction from Daniel Faggella, head of research at Emerj, in as little as five years, over 90% of pornography consumed by humans will be AI-generated. This is, according to him, because the content will expand beyond what we think as adult content today. 

“People will have access, not to ‘videos,’ but to one endless stream of dopamine-inducing content, calibrated in real time to precisely what they’re interested in or curious about,” said Faggella 

OpenAI’s response to inquiries about its stance on this matter suggests a cautious approach, with the company reaffirming its commitment to enforcing usage policies while acknowledging the importance of exploring conversations about sexuality in appropriate contexts.