> Microsoft collects speech, inking, and typing information
Is this the part where we expect to be able to use voice commands but object to the processing of voice data necessary to parse voice commands?
Siri has to listen to every word you ever say, send it to Apple and to the third party who actually does the processing for Siri, so she can tell if you said "hey Siri".
This data could be potentially aggregated or spied on, prior to being discarded after processing. Or you can just go without voice commands. Predictive keyboards and spell check work the same way, but with keystrokes.