> It seems like a short term solution to this might be to filter out any prompt content that looks like a policy file

This would significantly reduce the usefulness of the LLM, since programming is one of their main use cases. "Write a program that can parse this format" is a very common prompt.

Could be good for a non-programming, domain specific LLM though.

Good old-fashioned stop word detection and sentiment scoring could probably go a long way for those.

That doesn't really help with the general purpose LLMs, but that seems like a problem for those companies with deep pockets.