I've just read the several articles floating around, and here's what I think. It's a brief set of thoughts, but I'll publish a post to my blog on it.
They are not copying Google, directly.
From my statistical and analysis point of view, it looks like those using the toolbar and the browser combination sends information back to Bing, as part of CEIP or experience data. (Google does this, obviously). For strings which Bing is not sure about, it queries an authoritative source of information, e.g. Wikipaedia, Google, WolframAlpha. Google does this.
Search for 'diarea'. Google autocorrects this to diarrhoea, the correct spelling. Bing, on the other hand, does not, but pings Wikipeadia, then displays it first. However, if we define that search term, Bing does not show the autocorrection, but just does it, and shows the user the autocorrected term.
Both of the cases above indicate to me that there is some sort of internal lookup done, to provide the best answer. Is this rather smart? Yes. Google does that. Is this copying? no. If this was copying, then each test case scenario of authoritative and non-authoritative results would be exactly the same. Correlation does not mean causation.
Google's toolbar itself uses your experiences to generate its rankings etc.
On a different case, I spotted what looks like Google copying Bing's results:
http://www.google.com/search?q=site:.bing.com/search
Bing's robots/txt explicitly disallows /search. At time of writing, there are 25,400 results, up 400 from yesterday.