And their difficultly is a bonus in itself - if you want to hide your communications as much as possible.
Cost is not always the only metric to optimize. Even today, people try to use quantum entanglement to create communication channels where snooping is detectable.
EM already propagates across the known universe.
"And their difficultly is a bonus in itself - if you want to hide your communications as much as possible"
It is unlikely that such hiding is going to be more effective than existing spread-spectrum techniques, which can already put your communication arbitrarily close to the noise floor, and already exist.
Of course, if you're going to leave science behind you're then going to hypothesize what aliens may have done to somehow break through the noise floor so that's not an option, and, well, this is what I mean by pick one; either use science, or imagine whatever you like, but don't mix the two. There's no reason to believe that the noise floor is penetrable and lots of mathematical reasons to believe it very much isn't.
I meant it in the sense that EM will be attenuated by gas clouds and solid matter due to the high interaction rate of EM, while neutrino (or presumably dark matter) much less so.
So this excludes, for instance, the idea that there are dark matter civilizations using dark matter machines to send some sort of dark communication that we can't see.
It is not quite impossible to eliminate the possibility that there is some sort of dark radiation we could eventually create and detect. There's currently no particular reason to believe such a thing is possible, but it is not currently excluded necessarily. Whatever forces dark matter interacts with could have its own radiation. But it is difficult to imagine in what way this would be superior to EM-based communication, because, again, EM is just so perfect for communication that it's hard to beat.
It's the question people don't want to deal with because they're too busy being "gee whiz golly gosh" about the possibilities... in what way is this other communication method going to be better than EM? Dark matter communication would quite likely look in practice a lot like neutrino communication does today. It's already possible. We know that because we've built it. We don't call it a transmission method, though, because the bandwidth is soooo atrocious that it's reasonably approximated by "0 bits per second", so we just call it a "detector".