Fixed surveillance cameras might seem passive, but they really are typically more intrusive than this guy's project. That's what makes it so interesting to me. I totally understand how it'd be creepy to have someone physically walking around recording me, but in reality, a campus-wide set of "passive" surveillance cameras (say, at my workplace) will gather far more information about me than this guy could ever hope to record on his own.

Yeah I agree and with improvements in the technology it will get even worse. Cameras can already have pretty decent microphones and speech-to-text technology is pretty damn good so I wont be surprised if soon they start recording sound and analysing it in realtime. It wont cost all that much compared to a lot of other things.

However I was only trying to explain the behaviour of the people.

I think your assessment is right. Furthermore, I believe the intent of the video is to get people to question why these perceptions exist.