@AC Re: To inform or to entertain?
>If Google News chooses, for instance a Guardian headline over that of, say, The Times[...]
...then it's of little consequence, because grouped right next to it are all the other articles that (appear to) cover the same story. This is exactly why GN is my go-to for digging into a story: you're never more than a couple of clicks away from every single article they can find on the subject, and (this point is subjective, but hey) I've never gotten the sense that the "minority report" was omitted, or buried way down the list. It's pretty much always there among the first few, if it exists.
Yes, they are prioritising the sources, but AFAICT they're mainly doing so to appeal to your perceived preferences/bias based on their tracking data. (I've commentarded about this very thing in the last week or so.) This has its own risks if you are not alert to it, of course (essentially a positive feedback loop), but you can hardly blame them for prioritising the stuff they think you're most likely to want to read (and thus click on).
None of this prevents you digging deeper by yourself of course, as I'd expect a meat-sack journalist to do, but I'm just a lazy current affairs maven with limited time on my hands and I can't count the number of times I've read the first article about something, formed an opinion, and been very glad I also examined the other half-dozen versions at the top of GN before I started fulminating a load of underinformed bollocks all over the intertubes.
@Pete 2: Dabbsy already addressed your point about journos writing "for" GN, but it's always amusing to see when this has been unintentionally thwarted by the webmonkeys. For example, the scraped "summary" often contains the text of an advert that precedes the rubric, or the caption of the unnecessarily large and semantically irrelevant header image (well done on dodging that one, ElReg!).
My favourite at the moment is the Independent, where the scraped thumbnail is usually not from the article, but from the annoying tangentially-related-sometimes image slideshow they shove in half way down the rubric: witness an article about human rights abuses adorned with a pic of some grinning Japanese chaps emptying buckets of snow over their semi-naked selves (new season of Takeshi's Castle filming, I surmise). Great fun.
As to TFA, I think Dabbsy's overlooked the very virtue of robojournos: they have no bias apart from what is programmed into them. They also can't be bribed or coerced to spike or misrepresent a story, so actually they could make better investigative journos than any meatbag. Above all, they bring a standard of scepticism that no human can match: something they are told is either provably true, or it isn't. They can, if they're allowed, flag the article with "This could all be utter bullshit, by the way."
So it comes down to ownership, just as with the fleshies. If we could set up an open-source aggregator/miner bot with smart enough algos and somehow guarantee that it was neither tampered with by its masters nor could be hoodwinked into accepting lies as facts (now there's a challenge), I'd read it.