Google and Bing not only have their API-based translation, but also their websites which allow for one-off translation.
While these are intended for use by humans, this also allows for a webcrawler to automatically crawl the site and obtain translations of short phrases. If a human were to manually browse to a given provider's translation site and input a phrase to be translated, then stick it up on their blog, it seems like this would be a reasonable use.
However, what are the implications of crawling a translation service to obtain the results by proxy for the user? Sort of like TranslationParty but being ad-supported rather than ad-free?
The terms and conditions for Google, for example, seem pretty straightforward about fair use of their translation API, but use of the site interface seems unclear.