NATO on adversaries' language models

Many of us have been talking about this for a while, but here NATO has produced a concrete overview of how strategic adversaries can assemble LLMs into a formidable propaganda and influence machine. In this case it is a specific Reddit forum, but it is not hard to imagine this being done (and it surely already is) at a much larger scale.

How can an open democracy like ours defend itself against this? It is a question I hope we find good answers to fairly quickly.