That's the first I've heard of it. Seems interesting...until I hit the thought about who gets to do the work and where the data is going to come from. And then what happens when someone re-identifies an artist. And the large number of RAR files. But I'm "that guy," so don't necessarily go by me.
It's an interesting idea, of course. And it's something that might be nice to see happen as we have more comics--especially those anthology titles.
This might be more interesting as the GCD updates their system. I know an API was on the agenda at some point, meaning that a future comic-reading program could potentially say, "hey, there's more recent information about this comic available," then download, convert, and update your local metadata in the book. Maybe.
Though that could play havoc with MD5 hashes, for example, if I want to know if I'm missing any books between your library and mine. Unless the metadata also stores a hash of the images inside.
Ahem. Sorry, I get like that sometimes around software ideas. Occupational hazard.
To answer your original question, I don't think anybody here has mentioned it. I know that some books will have a text file inserted (that CBReader pops up in a window on opening the file) with a cleaned-up scrape of the GCD page. All the JVJ books start (or end) with a scan of Jim's index cards with most of the information--that's become a de facto standard just based on the number of his books now in circulation. But I don't think anything has become more sophisticated than that, so far.