HN Explorer

Ask HN: Do we need a process for excluding bodies of work from LLM

7 points by afinlayson ↗ HN
Do creators need defense from LLM scraping the internet? Create a robot-ai.txt file or include language in licenses that instruct artificial intelligence not to include certain content from large language models? Particularly for open-source software, webpages, and individuals who brand their work?

3 comments

[ 3.2 ms ] story [ 20.4 ms ] thread
nullish_signal
Much as """Open"""AI does not release Source Code, I am confused why Open-Source repos should be kept from AI training data?
afinlayson
Have you seen gpl 2? In theory that might cover this case. But there’s a lot of reasons open source people don’t want certain code being used without constraints but allow it to be open sourced.
pifm_guy
It works much the same way as you exclude the rest of the people in the world from a document... Don't publish it on the open internet...