This article makes a mistake that I've seen a hundred times before on Slashdot: confusing "the patent says...", and "the patentee intends to..."
I write software patents for a living. (I didn't write this one.) Let me describe how the patent drafting process goes.
A client comes to me with a simple invention - we'd like to do (A), (B), and (C) to achieve result (X). I talk to them at length about what (ABC) is, and what critically sets (ABC) apart from every similar example. I ask questions about how each of (A), (B), and (C) could be varied; what other elements (D), (E), and/or (F) could be added; and whether (ABC) could also be used for results (Y) or (Z).
And when I write up the patent application, EVERYTHING goes in there. (ABC) is described as the base invention, but all of the other material about (D), (E), (F), (X), (Y), and (Z) is also included as optional extensions or uses of (ABC).
Now, here's the critical thing: I haven't fully considered whether (D) is a desirable feature, or whether (Y) is a desirable result. My client doesn't even know, or says, "we don't really intend to implement (D) or do (X)." None of that is relevant. All that matters is: They are all logical, valid extensions of (ABC), so, typically, they all go in. Anything that could make the basic technique more valuable, appear more useful, or might more fully distinguish (ABC) over known techniques is helpful to add to the specification.
I read this patent the same way. The basic invention is: "Use a camera to count and identify people interacting with a device." Now, you can't just stop there - you haven't said what that information might be used for, and the patent office typically rejects applications that look like, "The technique is: Generate some data." So the patent discloses several uses of that information. That doesn't mean that Microsoft has any interest in using that technique - only that it's logically achievable from the basic techniques.
Look, we all agree that technology is neutral, right? For example, DRM has been *used* for lots of obnoxious purposes (including limiting fair-use rights), but the basic technology of DRM is neither good nor bad - it just is. The same principle applies here.