Google has added a rights filtering option to image search. It looks like this:
There are 5 options as you can see, based on Creative Commons licensing. If you choose an option, Google tends to use a smaller number of sites such as Flickr, Wikipedia and Wikimedia to pull back content. However, it's also using other resources as well, and this is where it gets a little confused. Google is looking for some sort of Creative Commons signing on a page, but what seems to be happening is that it assumes that anything it finds on that page is all Creative Commons, which isn't always the case. I found an example where a blogger has a CC licence for his material, but he'd used an image from elsewhere which wasn't his own. I traced it back to the original owner, who does, it's fair to say, have a CC licence, but which requires attributation back, and the blogger hadn't done this. So it would be far too easy to think the image was the bloggers when in actual fact it isn't.
It's a nice tool, no doubt about it, but Google has once again gone for a quick shiney fix rather than doing the job properly. If you're looking for images to re-use, you might want to take a look at search engines that work specifically in this area, and here's an element of one of my Pearltrees that does that: