Now it collects images and videos from vans with cameras for Street View (raised privacy concerns).
A lot of people get upset about Streetview, and I can understand why. But in my mind, and I'm a pretty paranoid/privacy-conscious person, I don't think Streetview is a big deal. As long as they're collecting data that's freely available already, I don't see how it's any different than spidering the Web.
It also collected 600GB of fragmentary data from open Wi-Fi networks recently ... I don't understand this one at all.
I don't know the details, but I have a suspicion of what happened. They were trying to collect MAC addresses, which has been done by other companies before. To do that, one of their engineers wrote a script that would take the first x bytes, say 64, from every packet. Or maybe they used tcpdump, which takes the first 68 bytes of each packet by default. They stored that without really thinking. But, it turns out, the tcp and ip headers are something like 20 bytes each, meaning 40 bytes of headers and approximately 20 bytes of payload. Oops?
The reason I'm not worried about this goof up is that it's Google themselves who admitted it was happening and announced that they'd be removing all the data they accidentally collected. If Google was collecting it for any reason other than an accident, why would they admit it and say they're getting rid of it?
As I said, I don't like Google or a lot of what they do, but these examples aren't something I'm worried about.