Rule Update: AI Generated Content

[Haste]

chatGPT says about this post:
Likelihood this was written by AI (or heavily AI-assisted):
15%–25%

1. Who asked?
   
2. What leads you to believe that ChatGPT itself is a good tool for analyzing whether or not something was made by ChatGPT?
   
3. What leads you to believe that this policy is going to be enforced by us copy-pasting forum messages into an AI model and just trusting whatever it claims?
   
4. For ChatGPT, 15-25% is a pretty low figure, and it's probably as low as it's going to go for any coherently-written post with mostly valid grammar and spelling. Which leads us back to point two.
   

We're trying to deal with blatant AI slop here, which you can recognize in very many ways. For example, AI-generated text might ramble on about following environmental laws and regulations in the Omicrons, and how proud the company is for following them so diligently. It'll use extremely generic sci-fi mumbo-jumbo language all the time while referring to basically no (Discovery) Freelancer lore or concepts.

Do:

• Use LLMs to help you draft roleplay posts and stories.
  
• Use LLMs to slightly improve some clunky, poorly-flowing sentences you wrote.
  
• Use LLMs to investigate what parts of the Discovery Freelancer timeline your character would have lived through, given a certain date of birth -- taking great care to verify the model's claims, as it may well hallucinate a dozen fake events.
  
• Use LLMs to generate the name of some imaginary park in the middle of an imaginary district on Planet Manhattan, if you're feeling a little uninspired.
  

Don't:

• Have an LLM literally write your roleplay for you.
  

It's really that easy.

---