this is a genius way for google to crowdsource future classifiers they may use on search results. the terms of the api say the original data is still owned by you, but it does not outline who owns the classification set generated from your data.

maybe i am wrong here. just seems like a good resource for them to use later.