Text files are generally more resistant to corruption, especially when parsed by a human, and more repairable if they do get corruption. (A one byte error probably won't matter in a text file; it could wreck parsing rigid structure.)

They more easily work with a lot of un-specialized tools, or tools specialized for text, which is a very general problem with very transferable learning.

Perhaps better to distinguish between "structured" and "unstructured" or "machine optimized" vs "human optimized".

A different beast at the end of the day, whatever name or distinction you use.

If you’re being pedantic, yes, but it’s a format that can be used by any number of tools that aren’t specific to systemd.