• Source:JND
Add as a preferred

AI research and deployment firm OpenAI released the first draft of its Model Spec, a document outlining the behaviour of its models in the OpenAI API and ChatGPT. Under its 'Don't respond with NSFW content', it says that the assistant shall not generate content which is Not Safe For Work (NSFW).

It includes content that may not be appropriate in a conversation and may include "erotica, extreme gore, slurs, and unsolicited profanity". However, OpenAI is planning to change this. According to the Model Spec draft, the company is exploring whether to allow users the capability to "generate NSFW content in age-appropriate contexts".

Also Read: OpenAI May Announce Its Search Engine One Day Before Google I/O 2024 | Here's What We Know

However, it added the company plans to better "understand the user and the societal expectations of model behaviour" in this space. "We're exploring whether we can responsibly provide the ability to generate NSFW content in age-appropriate contexts through the API and ChatGPT," the commentary noted.

The reason for exploring the change in the restrictions is to allow developers and users flexibility to use the services "as they see fit, so long as they comply with our usage policies", according to OpenAI. However, when Mashable reached out for comment, an OpenAI spokesperson said: "We have no intention to create AI-generated pornography".

As per the spokesperson, the company implements strong safeguards across its product line to curb deepfakes and prioritise the protection of children. "We also believe in the importance of carefully exploring conversations about sexuality in age-appropriate contexts."

Also Read: OpenAI Will Now Help Users Identify If Images Are Made By AI

For context, Model Spec is a document that lets users get a preliminary look at how an AI model is trained to respond to queries. The note also highlighted how ChatGPT and API may respond when asked to break laws, explore details around biological, radiological, nuclear and chemical threats and questions of explicit nature.