They're probably going to attempt to use the sensor data to get more data to form a mesh grid, so as to determine what point in a venue they're shooting.
For example, say there are four phones, in a square, and they all have color enabled, it would be possible to tell that they were all facing the middle of the square. From there, it'd be cool to stitch together the images and create a reverse panorama -- think of something like 'Streets' view, wherein you can 'enter' the scene and move around in it, from a variety of angles and perspectives.
At least, that's what I'd use it for.
