Churnalism is a brilliant idea - no surprise that it was originally one of Chris Lightfoot's. Basically, it allows you to determine how much of a given newspaper article was copied from which press release. There's a nice graphic visualisation, and a diff, so you can see precisely what was altered and what taken over in its entirety. It's right up there with Piggipedia and SukeyDating as a brilliant piece of geek activism.
However, here's something amusing. There's a basic API here; I chucked the text of the GSMA final press release from this year's MWC at it, and I was quite surprised at the results. The first article it extracted from Journalisted was none other than this piece in the Guardian from...February 2008. One consequence of churnalism is that your newspaper is likely to get repetitive.
As far as I can see, if there's anything missing here it's that the comparison is mostly the wrong way - having a newspaper article and wanting to know what vacuous NIB-fodder got regurgitated into it is a much more common use-case than having a press release and wanting to know which newspaper articles it got into. Actually, the latter use-case is far more likely if you're a PR and you're trying to measure how the talking-points are spreading. But once it has more press releases on file, it'll work better in that sense. And that's just a question of hoovering Businesswire, PRNewsWire etc up.