OpenAI released model safety guidance on Wednesday while acknowledging that it is investigating how to support the creation of content that is NSFW, or "not safe for work."
How to turn on/off NSFW content on Crushon AI (2024) | Turn on/off NSFW content on Crushon AI
The chatbot service provider's Model Spec is "a new document that specifies how we want our models to behave in the OpenAI API and ChatGPT." These guidelines are intended to provide machine learning researchers and data labelers with recommendations on how to fine-tune models using a technique called reinforcement learning from human feedback (RLHF).
For example, the model specification states that generative AI assistant applications "shall not provide content that is Not Safe For Work (NSFW): Content that would not be appropriate in a conversation in a professional environment, which may include eroticism, extreme gore, profanity, and unwanted profanity ."
Meanwhile, OpenAI says it's considering just the opposite.