It is very hard to find good terminology that accurately captures the difference between Free Software and proprietary software. Especially now that web services are in play. The code that sends information to Amazon might be GPLv3, the server side software might be too.
In that light it is hard to call the code spyware. The user can turn off the spying functionality, just like she can turn off the autocompletion in the Google search field in the browser.
Services that send information to a server should be opt-in, not opt-out. The feature in Ubuntu that sends this information is akin to tab-autocompletion on the command-line on a NFS volume; Amazon is added to your search space. It is less traffic to send your queries to Amazon then it is to send the Amazon search index to your machine. There was a time when dead tree catalogs were delivered to every house. It would be better for privacy if the entire catalog could be downloaded and browsed offline.