Human authorities vs machine generated wisdom

In my last newspaper article in the NZZ I end up with the question

"ob nicht Websites, die allen offenstehen und über
ausgeklügelte Qualitätskontrollen für die Fakten
verfügen, besser als einzelne Spezialisten "wissen
", was wichtig und aktuell ist.
"

That is, I question whether a human expert ought be replaced by some kind of an artificial system delivering a synthesis and a metrics of how trustworthy a bit of information is. This can cover the synthesis, but could go down to the single facts upon which such an synthesis is based.

Rod in his Iphylo blog had an interesting comment about this issue, well worth reading, including Matt Cockerill's post.

I would shy away from saying, that no humans are needed, but ponder rather the idea that for knowledge based on data on Internet-accessible sources and predefined criteria, a machine might fare better.

This is certainly a hot issue regarding the measuring a scientists contribution and needs with new models of communication, such as plos one, blogs, databases the attention of those needing to judge scientists for their promotion.