The only way it could possibly censor sensitive information is if it captured it in the first place and then determined that it was in a sensitive category and then censored it. Recall still has to capture it first to make that determination.

I don’t understand why this isn’t everyone’s immediate thought after hearing Microsoft say their system would censor sensitive information. How could it possibly know what to censor without reading it first? Of course it’s going to invade your privacy, and then maybe they’ll selectively delete some of it when you ask them to.

I wouldn’t be surprised if it all gets uploaded to cloud storage first, and then the “sensitive” stuff gets deleted from the local storage only.