Establishing an etiquette for LLM use on Libera.Chat
Greetings fellow humans!
For better or for worse, controversially perhaps, we are now immersed in an online landscape laden with LLMs.
This may be a fact that brings you great joy. However, not everyone feels the same way about LLMs. Some folks dislike them for privacy or ethical reasons. Some people just don’t feel good about talking to LLMs without knowing. We want folks in all camps to feel welcome on Libera.Chat.
We believe that the best way to make this happen is to ensure that people are not caught unaware. As such, building on the principles of our public logging policy, staff would appreciate cooperation with these simple guidelines:
- LLMs are allowed on Libera.Chat. They may both take input from Libera.Chat and output responses to Libera.Chat.
- Training LLMs on channel content or logs is subject to the public logging policy and therefore requires permission before that content can be used for training.
- As soon as possible, people should be made aware if they are interacting with, or their activity is being seen by, a LLM. Consider using line prefixes, channel topics, or channel entry messages.
- People who operate interactive LLM scripts or bots in channels they don’t run must obtain permission from the channel founders and operators.
- LLM scripts and bots are not exempt from our network policies. People running LLM scripts and bots are responsible for what they output.
- Other people may be responsible for what a LLM script or bot outputs based on their prompts.
While these guidelines are not final, nor enshrined as network policy with exception of training LLMs on channel content, we do have a broad prohibition on antisocial behaviour. We encourage folks to consider whether their application of LLMs is reasonably prosocial or antisocial.